You can subscribe to this list here.
2004 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(3) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2005 |
Jan
|
Feb
|
Mar
(1) |
Apr
(30) |
May
(5) |
Jun
(4) |
Jul
(13) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2006 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(4) |
Sep
|
Oct
|
Nov
|
Dec
|
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Catherine B. <BA...@ge...> - 2007-06-11 17:23:31
|
Dear fellow MGED member: Several members of the Microarray Gene Expression Data Society (MGED; www.mged.org) are applying to the NIH for funding for continued software and standards development. Your feedback is the most valuable tool we have for ensuring that our proposal is timely and serves the needs of the research community that generates and uses high-throughput biomedical data. Could you take a moment to answer a few questions about our MGED standards, file formats and software? It will take less than 10 minutes to fill out this survey. Your input will help us immensely in proposing work that will be of value to your research. http://www.surveymonkey.com/s.aspx?sm=RT_2f_2bcjSc3YjyZ7lhBZmBsA_3d_3d Many thanks, Cathy Catherine A. Ball Stanford Microarray Database Department of Biochemistry Stanford University School of Medicine Stanford, CA 94305-5307 USA e-mail: ba...@ge... www: genome-www.stanford.edu phone: (650)724-3028 fax: (650)724-3701 |
From: Thorsten H. <he...@em...> - 2006-08-30 14:41:40
|
Hi Delphine and Eric, within the last year we have set up a cross-species expression =20 pattern database. We have imported drosophila and zebrafish data so =20 far and will do Medaka and mouse soon. The schema we have developed =20 should be Misfishie compliant (I hope) and I can send it to you if =20 you like. Our database is separated into a repository (normalized, all data) at =20= EMBL Heidelberg and a Warehouse (only data necessary for queries, =20 fast) at the EBI in Hinxton UK (ArrayExpress group). We have =20 developed a XML format to exchange data, but this far from a GELE-ML=20 (gene expression localization experiment-ML). I have a master =20 student, who has just started to do the first steps towards a GELE-=20 ML, but not much to present yet. Anyway I think this needs to be an =20 effort of not just one group but rather a consortium as it will be =20 used by many groups... Last month I presented our project on a meeting in Roscoff, where I =20 met Patrick Lemaire (ANISEED). We discussed about an in situ data =20 exchange format and agreed that it would be important to have a small =20= meeting of people who work on in-situ and similar databases to start =20 with it. Patrick suggested beginning of next year, if I remember =20 correctly and will announce it and ask around for interested =20 researchers, who would like to participate. Regards, Thorsten On Aug 30, 2006, at 10:45 AM, Delphine Dauga wrote: > Ok, thanks. > I think, we have to adapt our database to MISFISHIE because of the =20 > precision of the information. > We have already elaborated a dtd but specific to ANISEED (website: =20 > http://crfb.univ-mrs.fr/aniseed/) and it will be judicious to =20 > generalize to the others database. So I try to make a "mix" between =20= > the Minimum Information Specification For In Situ Hybridization and =20= > Immunohistochemistry Experiments and that we have already done. > So, if you want to collaborate... > I will keep you informed of my work. > > Regards, > Delphine. > > Eric Deutsch a =E9crit : >> Dear Delpine, we have worked on a minimum information =20 >> specification (the >> manuscript is currently under review) but have not yet worked in =20 >> earnest >> on an official data model or DTD. We have some sample database =20 >> schemas >> and the like. Last I heard Thorsten (CC'ed here) was working on a =20= >> data >> model; perhaps he would chime in if he has something ready to share? >> >> Regards, >> Eric >> >> >> >>> From: Delphine Dauga [mailto:da...@ib...] >>> >>> Dear Eric Deutsch, >>> >>> I am working on ANISEED (Ascidian Network for In Situ Expression and >>> Embryological Data) and I wondered whether an exchange format =20 >>> already >>> exist. I found MISFISHIE but I have the feeling that the project is >>> "suspended" (no mail on the mailing list since 26/07/05). Is someone >>> currently on the project? >>> I try to elaborate a dtd for this type of information but maybe it >>> already exist. If I can discuss with someone about that, it will be >>> >> more >> >>> effective. >>> >>> Thank you for your answer... >>> Best regards >>> Delphine Dauga >>> >> >> >> > > --=20 > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > > Delphine DAUGA, Aniseed Curator > Equipe Lemaire > IBDML - CNRS, Marseille - France > Campus de Luminy, Case 907 > F-13288 Marseille cedex 9 > Tel: +33 (04).91.82.94.18 > Fax: +33 (04).91.82.09.82 > da...@ib... > http://crfb.univ-mrs.fr/ciona/lemaire/ > http://crfb.univ-mrs.fr/aniseed/ > > > |
From: Delphine D. <da...@ib...> - 2006-08-30 08:44:10
|
Ok, thanks. I think, we have to adapt our database to MISFISHIE because of the=20 precision of the information. We have already elaborated a dtd but specific to ANISEED (website:=20 http://crfb.univ-mrs.fr/aniseed/) and it will be judicious to generalize=20 to the others database. So I try to make a "mix" between the Minimum=20 Information Specification For In Situ Hybridization and=20 Immunohistochemistry Experiments and that we have already done. So, if you want to collaborate... I will keep you informed of my work. Regards, Delphine. Eric Deutsch a =E9crit : > Dear Delpine, we have worked on a minimum information specification (th= e > manuscript is currently under review) but have not yet worked in earnes= t > on an official data model or DTD. We have some sample database schemas > and the like. Last I heard Thorsten (CC'ed here) was working on a data > model; perhaps he would chime in if he has something ready to share? > > Regards, > Eric > > > =20 >> From: Delphine Dauga [mailto:da...@ib...] >> >> Dear Eric Deutsch, >> >> I am working on ANISEED (Ascidian Network for In Situ Expression and >> Embryological Data) and I wondered whether an exchange format already >> exist. I found MISFISHIE but I have the feeling that the project is >> "suspended" (no mail on the mailing list since 26/07/05). Is someone >> currently on the project? >> I try to elaborate a dtd for this type of information but maybe it >> already exist. If I can discuss with someone about that, it will be >> =20 > more > =20 >> effective. >> >> Thank you for your answer... >> Best regards >> Delphine Dauga >> =20 > > > =20 --=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Delphine DAUGA, Aniseed Curator Equipe Lemaire IBDML - CNRS, Marseille - France Campus de Luminy, Case 907 F-13288 Marseille cedex 9 Tel: +33 (04).91.82.94.18 Fax: +33 (04).91.82.09.82 da...@ib... http://crfb.univ-mrs.fr/ciona/lemaire/ http://crfb.univ-mrs.fr/aniseed/ |
From: Eric D. <ede...@sy...> - 2006-08-30 05:41:20
|
Dear Delpine, we have worked on a minimum information specification (the manuscript is currently under review) but have not yet worked in earnest on an official data model or DTD. We have some sample database schemas and the like. Last I heard Thorsten (CC'ed here) was working on a data model; perhaps he would chime in if he has something ready to share? Regards, Eric > From: Delphine Dauga [mailto:da...@ib...] >=20 > Dear Eric Deutsch, >=20 > I am working on ANISEED (Ascidian Network for In Situ Expression and > Embryological Data) and I wondered whether an exchange format already > exist. I found MISFISHIE but I have the feeling that the project is > "suspended" (no mail on the mailing list since 26/07/05). Is someone > currently on the project? > I try to elaborate a dtd for this type of information but maybe it > already exist. If I can discuss with someone about that, it will be more > effective. >=20 > Thank you for your answer... > Best regards > Delphine Dauga*///// > /* >=20 > -- > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >=20 > Delphine DAUGA, Aniseed Curator > Equipe Lemaire > IBDML - CNRS, Marseille - France > Campus de Luminy, Case 907 > F-13288 Marseille cedex 9 > Tel: +33 (04).91.82.94.18 > Fax: +33 (04).91.82.09.82 > da...@ib... > http://crfb.univ-mrs.fr/ciona/lemaire/ > http://crfb.univ-mrs.fr/aniseed/ >=20 |
From: Delphine D. <da...@ib...> - 2006-08-29 10:19:47
|
Hi everybody, I am discovering MISFISHIE and I wondered whether there were already a=20 DTD=85 I don't find anything about this. Best regards, Delphine Dauga --=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Delphine DAUGA, Aniseed Curator Equipe Lemaire IBDML - CNRS, Marseille - France Campus de Luminy, Case 907 F-13288 Marseille cedex 9 Tel: +33 (04).91.82.94.18 Fax: +33 (04).91.82.09.82 da...@ib... http://crfb.univ-mrs.fr/ciona/lemaire/ http://crfb.univ-mrs.fr/aniseed/ |
From: Eric D. <ede...@sy...> - 2005-07-26 20:40:54
|
Hi everyone, Thank you to all who provided feedback to previous drafts. I have addressed he latest comments from Martin and Steve Bova (and a few minor things from others) and merged everything into another draft, and here is the result: http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v16.doc Please read it over and send back any comments to the email list: mge...@li... If you do not belong to this list, your may join or view the archives or membership at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ I encourage you to forward this message to additional people whom you think should see the manuscript but may not be on these distribution lists. Note that you can view the current members list at the SourceForge URL above. I will also be sending this manuscript to a few more members of the community not on these lists, seeking input from them. If you would like to suggest additional people and would rather have me send them an email, let me know. Regards, Eric ---------------------------------------------------------------------- Eric Deutsch email: ede...@sy... Institute for Systems Biology Voice: (206) 732-1397 1441 N 34th St FAX: (206) 732-1260 Seattle, WA 98103-8904 http://www.systemsbiology.org/ |
From: Eric D. <ede...@sy...> - 2005-07-19 23:11:23
|
Hi Martin, thank you for your response. I reply below in-line... =20 _____ =20 From: Martin Ringwald [mailto:rin...@in...]=20 Sent: Wednesday, July 13, 2005 4:18 PM To: Eric Deutsch Cc: mge...@li...; rin...@in... Subject: Re: Fwd: [Mged-misfishie] MISFISHIE =20 =20 Hi Martin and Duncan, thank you for posting your comments on the manuscript here. I spoke at length with Larry about how to address your concerns. We have made changes in response to nearly all of them, and would be grateful if you would reread the latest draft and let us know whether you are happy with the revision, or if you have further comments. Although there have been no other responses to your comments, I highly encourage others out there to add any comments they wish. =20 Eric and Larry, =20 Thanks for addressing our concerns. And sorry for not responding sooner. I am currently away from my desk and have only sporadic access to my email. My comments are inserted below. =20 =20 > Experimental Design: >=20 > Total number of hybridizations/stains performed in the experiment: a hybridization/stain > is defined as an assay of a single specimen with a single reporter. Thus, the result of a > tissue microarray consisting of a 10 x 10 array of different tissues would be counted as > 100 immunostains. If replicates or reruns are a component of the experimental design, > provide details that should include number of replicates per tissue, per reporter, etc. >=20 > This might make sense for tissue arrays but I don't think it can be readily applied to > conventional RNA in situ and immunohistochemistry studies (as a minimal requirement). > Of course, people make replicate experiments but most of the time they don't do this in a > systematic fashion and regular papers rarely, if ever, provide this type of information. > (Yes, you could argue that in these cases replicates are not a component of the > experimental design.) We didn't understand why you don't think these two requirements: =20 - total # of stains reported on - *IF* there are replicates as part of the of the design, tell us are applicable to conventional ISH and IHC. We did find several papers that mentioned replicate information ("each antibody assayed with 3 - 5 separate tissues"). If the papers were silent on replicates, we assumed there were none. Would it be a bad thing to force authors to state or imply "we performed no replicates"? Duncan commented: I think authors should be required to describe the number of replicates in their data. As Martin says, people often don't do a strict series of replicates, but I would say that they should be required to show that they have done enough replicates to substantiate the results. MISFISHIE currently doesn't even require that they show that they have done enough replicates. It only requires that they reveal what they did. It would be up to a reviewer (or the reader) to decide if it was "enough". In our view, this requirement should provoke authors to write 1 sentence in their manuscript saying what the total assay count was and how replicates are done if they are done at all. This does not seem burdensome. If they don't do replicates, that's fine as far as MISFISHIE goes. If the reviewer is happy about no replicates, then all is well. =20 =20 Considering all the arguments, this section is probably fine as is. =20 What, exactly, is the MISFISHIE definition of an assay? Would the tissue microarray described above correspond to 1 assay or to 100 assays? For a conventional ISH or IHC, would each stain correspond to an assay, or could an assay consist of, or be a representative of, many stains, including stains from replicate experiments? Maybe I simply misunderstood the relationship between stains, assays, and reported results. =20 Stains and assays are the same thing. We use assay as a more generic term or hybridization or stain. I have tried to clear this up a little. MISFISHIE now reads: =20 Total number of assays performed in the experiment: an assay is defined as one instance of a hybridization/stain of a single specimen with a single reporter. Thus, the result of a tissue microarray consisting of a 10 x 10 array of tissues would be counted as 100 assays. If replicates or reruns are a component of the experimental design, provide details that should include number of replicates per tissue, per reporter, etc. =20 =20 We don't feel strongly about this point. We'd like to ask you what you suggest we write here? Strike it completely? Reword? Please suggest what you would like to see in the manuscript regarding this. > Biomaterials and Treatments: >=20 > Manner of preparation of the specimens for the study. Information required includes the > nature of the samples: e.g. whole tissue, tissue sections, thickness of sections, whole > cells, =20 > or sections of cells; manner in which the specimens were prepared for the experiments, > e.g. fixation with type of fixative and duration of fixation vs. fresh, non-fixed, non-frozen > specimens or frozen specimens, sections mounted on slides versus sections floating in > reagents, nature of the slides on which sections were mounted; and the protocols used. >=20 > It would be good to have this data but I don't think information such as the nature of the > slides on which sections were mounted should be mandatory or is even that useful. We will strike the "nature of the slides". We will also adjust the punctuation a little to emphasize that there are only 3 requirements intended here: 1. nature of the samples 2. manner in which the specimens were prepared for the experiments 3. protocols used Everything else in the paragraph is examples of the sorts of things that should be considered for inclusion. We are not intending to require all of the examples provided, but rather only require that the paper describe the three above points so that the experiment can be properly assessed and replicated (to the satisfaction of the reviewer). Do you disagree with any of these three requirements? Or was it just unclear that the examples we provided were just examples of things that one could mention? I have tried to make this clearer in the revised text. =20 =20 I agree with these three requirements. The revised text is more clear. =20 Great! =20 =20 > Details on how the specimens were stored until use may optionally be provided. For > example, if frozen samples were used, information should be provided regarding storage > temperature and duration of storage. >=20 > I realize that this is optional. Nonetheless, this type of information is, in our experience, > never reported. Mentioning this optional requirement could reinforce a reviewer's > perception that MISFISHIE is overkill. We understand your point here and debated it here at length. We note that we have seen some papers where this information was revealed. The highlights of the debate may be summarized as: 1. Is there a role for optional elements in MISFISHIE? Would it be better to remove anything optional and just have bare minimum? Or would it be advantageous to specify some optional "best practices" that we encourage authors to provide if they wish, but not lay it out as a requirement? Maybe I'll throw this point by itself out to the list to see what reactions are. =20 2. We were inclined to leave it in, and see what our reviewers think. 3. Larry suggested raising this point in the submission cover letter. Anyway, I have removed the mention of storage parameters. =20 In general, I think it is a good idea to specify some optional best practises. =20 =20 =20 =20 > Reporters (probes or antibodies): >=20 > Protocol(s) for how the reporters were designed and produced or the source from which > they were obtained. For GFP-like experiments[HC1], the promoter sequence should be specified. >=20 > o For reporters purchased from a company, the company name and > catalogue number should be provided, as well as the web site, if available, > that provides details of the specifications. In addition, key aspects in the > specifications should be repeated since catalogue numbers and company > literature may not be available in the future. >=20 > All these points are well taken. However, given that it is already a requirement to provide > the sequence of the promoter, this is too much additional detail as far as minimal > requirements are concerned. Okay, you're right, this could be quite burdensome. The promoter part will be moved elsewhere, it doesn't belong in this passage. We propose to remove "as well as the web site" and everything after that. Therefore, the new text would just require company names and catalog numbers when reporters are purchased from companies. Would you agree that this should be minimum, or do you think this is still too much? This is now the same as the requirement on your GXD gxd_submission_guidelines.shtml page (referenced at bottom). =20 =20 As far as commercial antibodies are concerned, I think it is a good idea to ask for company names and catalog numbers. For a nucleotide probe, providing the sequence of the probe should be sufficient.=20 =20 Okay =20 =20 > Both positive and negative measurements of staining relevant to the experiment > should be reported. It is quite useful to provide negative expression results; it is > understood that a negative result is actually an upper limit to the expression level, > where the limit is usually not well known. >=20 > Sounds good. >=20 > For example: > Luminal epithelial cell: present > Basal epithelial cell: absent > etc. > or: > Luminal epithelial cell: 90% present, 10% equivocal, 0% absent > Basal epithelial cell: 0% present, 20% equivocal, 80% absent > etc. >=20 > This detail in data annotation would be ideal and, obviously, it is just an example. > Nonetheless it is worth pointing out that it makes a huge difference if one has to annotate > a section of a part of a tissue or of a whole organism. The specification of the required > detail of annotations must take this into account, i.e. the amount of work must be kept > reasonable for the task at hand. Other issues to consider here are space limitations in > journal publications and readability of the manuscript. Electronic data submissions to > pertinent public databases have their place here but, again, the requirements must be > kept reasonable. Although we don't present it here in tabular form in the example, we'll clarify that it would best be presented in tabular form, either within the manuscript or as supplemental material depending on the scope of the data. Duncan further commented: I also still feel that we should be careful to distinguish the requirements for results from the interpretation and annotation of these results. In particular, I think it is unwise to require that people annotate all of their data. Many in situ experiments are targeted at specific organs, but incidentally give results for other tissues. A requirement for complete annotation runs the danger of pressing people to make inexpert annotations of the latter in order to push through publication. =20 =20 If a study targets a specific organ and does not annotate the results for other tissues, I think it would be very helpful to state this fact explicitly in the paper. =20 =20 I think we made it clear that we aren't requiring totally complete annotation of all possible visible structural units, but rather just of the structural units of interest to the paper, i.e. if positive expression for one reporter is listed for a structural unit, then that structural unit should be characterized for all other reporters, too.=20 =20 Yes, this is what I would like to see, too. This is particularly true for your example below: =20 =20 Do I correctly interpret your comments to reflect that I could say that "CD1 is present in cell type A, CD2 is present in cell type B, and CD3 is present in cell types B and C" and that by itself should be entirely sufficient as a minimum. i.e. I have not explicitly provided any information on whether CD1 is expressed in cell type B and that is fine. As previously written/envisioned we would require a table indicating calls for CD1,CD2,CD3 in cell types A,B,C, i.e. 9 measurements. But you feel this is too much? This is a critical point here. If we relax this requirement, then there's essentially no improvement of the status quo and negative results won't get reported. With some regret, I have added a sentence indicating that nulls are allowed. =20 =20 My opinion regarding your example is: If the objective of a study is to examine / compare the expression of CD1, CD2, and CD3 in three different cell types, information on whether CD1 is expressed in cell type B should be provided explicitly. =20 Good, it sounds like we are in agreement. After the last revision, I inserted the weasel clause: =20 If some structural units cannot be characterized for some reporters, corresponding calls may be null. =20 Opinions for leaving it in or striking? =20 =20 > Protocol for the characterization and information about the basic technique for > characterizing the assays. For example, this information may include how many > observers performed the characterizations, assessment of inter-observer variability, > whether the characterizations were performed from the images themselves or visually > through instrument, any exceptions or assumptions made in characterizing the data, etc. >=20 > This sounds to me like maximum rather than minimum information. Further, in most =20 > cases the real number of observers is 1 (or maybe 2 max) and there is no assessment of > inter-observer variability. >=20 > One general problem with the manuscript might be that it tries to describe a best practices > effort by the NIDDK consortium, a very valid goal in and of itself and, at the same time, > suggests these best practices as the minimum standard for the whole community. > Researchers doing these types of studies might be more receptive to the paper if it is > more pitched as a best practices effort. We propose to strike the inter-observer part completely, and rephrase the "protocol for characterization" as an optional best-practice, but not a requirement, as with some of the other protocols. =20 =20 Sounds good. =20 Thanks also for your other comments and explanations. =20 Martin =20 =20 |
From: Eric D. <ede...@sy...> - 2005-07-19 21:14:10
|
Dear Steve, thank you very much for you comments. I will address each below inline. Please do respond and/or elaborate when you have time. > From: mge...@li... [mailto:mged-misfishie- > ad...@li...] On Behalf Of G. Steven Bova >=20 > Dear Eric, Larry, Alvin, Christian, John and others on the MISFISHIE > list that I don't know. Thanks for working on this, it could lead to > badly needed improvement in published immunostain, FISH, and related > data. I joined the list only recently and have a few short comments > that I can elaborate on in another email (I'll be back from vacation > next week) if you find any of them helpful. >=20 > These are in no particular order: >=20 > -In our own tissue microarray/antibody staining work, the thing we find > most difficult using the literature right now in attempts to > validate/replicate studies are the lack of specific replicable positive > and negative control data, and the lack of specificity on the exact > antibody clone used. For example, there are several papers studying > MDM2 expression in prostate cancer, but none report what they used as a > positive or negative control (there is a lung cancer cell line commonly The last item in the "Staining" section specifies the requirement for: Protocols for assay controls: the nature of negative tissue and reporter controls and optionally specificity reporter controls, such as competitive inhibition with either purified protein or peptide in immunohistochemistry Does this seem sufficient? If you would like to make a suggestion for a rewording or change in this section, please do so. Maybe we could make positive controls more explicit. We should be careful about being too demanding. > used for this), and in addition, several suppliers provide more than one > MDM2 antibody, so while all publications include the name of the > supplier, a fair number don't specify the clone, making truly impossible > to replicate without getting in touch with the authors, who often don't > remember. I think the checklist would be a good place to prompt > authors/reviewers to remember to include the supplier_clone_name. The The second major requirement in the Reporters section is: If known, the full sequence of the probe, or clone identifiers of the antibody. and in addition, for purchased antibodies we say: For reporters purchased from a company, the company name and catalogue number should be provided. Does this seem like we have covered this item pretty well, or do you suggest a change here? > specific positive and negative controls used (and how prepared) should > also be specifically requested...and I think it would be good to > encourage people to use controls that are freely available...otherwise > the chain leading to replication or refutation will be tenuous. The negative control protocol is already a pretty explicit requirement, but you're right that the positive control is implied at best. You would advocate specifically mentioning the positive control? Many researchers may not be doing positive controls routinely. > -For each tissue studied, the anatomic source beyond species should be > specified...ie, the liver, adrenal, lung, right lower lobe of lung, etc. > I looked through the specification several times, I must be missing > this, but I couldn't find where the specification asks for this. This is what we're calling a structural unit in the Characterization section. We say: Structural units could be an organ, tissue, cell, subcellular component, etc. Note that only the structural units relevant to the experiment need to be listed and characterized. It is not necessary to list (and characterize) structural units visible in the assays or slides but not relevant to the experiment or report. If a specific organ is prepared for the assays, that should be included in the Biomaterials section. But you're right, that it's not explicit about that. We could add to the Biomaterial section to provide the anatomical part or parts if the specimens are not whole. I wonder if this would lead to some duplication of information. > -Antibodies have Class/Subclass and Type/Subtypes categorizations. > They aren't that complex and could be defined specifically in the > checklist or as a reference table to the checklist to help the > author/reviewer comply. I think the term "isotype" used currently in > the specfication refers to Class, but I'm not sure. This sounds like this may be asking too much of a minimum requirement spec. What do others think? > -There may be a bit of a conflict related to images in that those who > are reporting results from reading images directly through a microscope > or from a live video image, if they comply, could easily provide images > that are of regions that were not the same as those used in their > scoring. Based on the following study that we did a while ago: > http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=3DRetrieve&db=3Dpubmed&= dop t=3D > Abstract&list_uids=3D11331959&query_hl=3D2 > I think there is strong enough evidence to encourage those publishing > immunostaining or FISH results to always define and store the set of > images to be scored first, then score the electronically stored > image.... in this paper, we were struck by the many advantages (in terms > of replication, decreased intraobserver and interobserver variability, > etc) of doing these studies this way, and only this way. You have an > undefined kettle of fish otherwise. I agree that this may well be a good practice, but I am very hesitant to add this as a requirement for one main reason: the goal for MISFISHIE is to specify a standard for disclosure of protocols, not to require that certain protocols be performed. We current say at the end of the Characterization section: Optionally as a best practice, the protocol for the characterization and information about the basic technique for characterizing the assays. For example, this information may include how many observers performed the characterizations, whether the characterizations were performed from the images themselves or visually through instrument, any exceptions or assumptions made in characterizing the data, etc. This used to be a formal requirement, but was reduced to an optional best practice because it was thought to be asking too much. I would be happy to add in a sentence noting and referencing the result of the paper you mention above here in the manuscript. But it would only be presented as a note in passing about a good best practice. > -Another point I think that would be worth making in the paper is that > there is still a lot of room for research in defining scientific best > practices for doing and reporting these types of studies....it might be > worth adding this to your statement that the group will be working to > improve the standard over time. Yes, this sounds like a good point to make. I will add this. > Let me know if any of this is helpful, the small public library I'm in > in Blue Hill Maine is closing, so I have to stop here. This has been very helpful. Please do respond to my responses and forward any other comments you may have. Thank you, Eric > All the best, >=20 > Steve Bova >=20 >=20 >=20 >=20 > G. Steven Bova M.D. > Asst Prof. of Pathology > Asst Prof of Health Sci Informatics, Genetic Medicine, Oncology, > Urology > Department of Pathology > Johns Hopkins University School of Medicine > Carnegie Building Rm 628 > 600 North Wolfe St. > Baltimore, MD 21287-6417 >=20 > Email: gb...@jh... > Fax: 410-614-7620 > Voicemail: 410-614-5957 >=20 >=20 > ------------------------------------------------------- > SF.Net email is sponsored by: Discover Easy Linux Migration Strategies > from IBM. Find simple to follow Roadmaps, straightforward articles, > informative Webcasts and more! Get everything you need to get up to > speed, fast. = http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclick > _______________________________________________ > Mged-misfishie mailing list > Mge...@li... > https://lists.sourceforge.net/lists/listinfo/mged-misfishie |
From: Martin R. <rin...@in...> - 2005-07-13 23:14:34
|
>> >> >>Hi Martin and Duncan, thank you for posting your comments on the >>manuscript here. I spoke at length with Larry about how to address >>your concerns. We have made changes in response to nearly all of >>them, and would be grateful if you would reread the latest draft >>and let us know whether you are happy with the revision, or if you >>have further comments. Although there have been no other responses >>to your comments, I highly encourage others out there to add any >>comments they wish. Eric and Larry, Thanks for addressing our concerns. And sorry for not responding sooner. I am currently away from my desk and have only sporadic access to my email. My comments are inserted below. >> >> >> > Experimental Design: >> >> > >> >> > Total number of hybridizations/stains performed in the >>experiment: a hybridization/stain >> >> > is defined as an assay of a single specimen with a single >>reporter. Thus, the result of a >> >> > tissue microarray consisting of a 10 x 10 array of different >>tissues would be counted as >> >> > 100 immunostains. If replicates or reruns are a component of the >>experimental design, >> >> > provide details that should include number of replicates per >>tissue, per reporter, etc. >> >> > >> >> > This might make sense for tissue arrays but I don't think it can >>be readily applied to >> >> > conventional RNA in situ and immunohistochemistry studies (as a >>minimal requirement). >> >> > Of course, people make replicate experiments but most of the >>time they don't do this in a >> >> > systematic fashion and regular papers rarely, if ever, provide >>this type of information. >> >> > (Yes, you could argue that in these cases replicates are not a >>component of the >> >> > experimental design.) >> >>We didn't understand why you don't think these two requirements: >> >> - total # of stains reported on >> >> - *IF* there are replicates as part of the of the design, tell us >> >>are applicable to conventional ISH and IHC. We did find several >>papers that mentioned replicate information ("each antibody assayed >>with 3 - 5 separate tissues"). If the papers were silent on >>replicates, we assumed there were none. Would it be a bad thing to >>force authors to state or imply "we performed no replicates"? >> >>Duncan commented: >> >>I think authors should be required to describe the number of >>replicates in their data. As Martin says, people often don't do a >>strict series of replicates, but I would say that they should be >>required to show that they have done enough replicates to >>substantiate the results. >> >>MISFISHIE currently doesn't even require that they show that they >>have done enough replicates. It only requires that they reveal >>what they did. It would be up to a reviewer (or the reader) to >>decide if it was "enough". >> >>In our view, this requirement should provoke authors to write 1 >>sentence in their manuscript saying what the total assay count was >>and how replicates are done if they are done at all. This does not >>seem burdensome. If they don't do replicates, that's fine as far >>as MISFISHIE goes. If the reviewer is happy about no replicates, >>then all is well. Considering all the arguments, this section is probably fine as is. What, exactly, is the MISFISHIE definition of an assay? Would the tissue microarray described above correspond to 1 assay or to 100 assays? For a conventional ISH or IHC, would each stain correspond to an assay, or could an assay consist of, or be a representative of, many stains, including stains from replicate experiments? Maybe I simply misunderstood the relationship between stains, assays, and reported results. >> >>We don't feel strongly about this point. We'd like to ask you what >>you suggest we write here? Strike it completely? Reword? Please >>suggest what you would like to see in the manuscript regarding this. >> >> > Biomaterials and Treatments: >> >> > >> >> > Manner of preparation of the specimens for the study. >>Information required includes the >> >> > nature of the samples: e.g. whole tissue, tissue sections, >>thickness of sections, whole > cells, >> >> > or sections of cells; manner in which the specimens were >>prepared for the experiments, >> >> > e.g. fixation with type of fixative and duration of fixation vs. >>fresh, non-fixed, non-frozen >> >> > specimens or frozen specimens, sections mounted on slides versus >>sections floating in >> >> > reagents, nature of the slides on which sections were mounted; >>and the protocols used. >> >> > >> >> > It would be good to have this data but I don't think information >>such as the nature of the >> >> > slides on which sections were mounted should be mandatory or is >>even that useful. >> >>We will strike the "nature of the slides". We will also adjust the >>punctuation a little to emphasize that there are only 3 >>requirements intended here: >> >>1. nature of the samples >> >>2. manner in which the specimens were prepared for the experiments >> >>3. protocols used >> >>Everything else in the paragraph is examples of the sorts of things >>that should be considered for inclusion. We are not intending to >>require all of the examples provided, but rather only require that >>the paper describe the three above points so that the experiment >>can be properly assessed and replicated (to the satisfaction of the >>reviewer). >> >>Do you disagree with any of these three requirements? Or was it >>just unclear that the examples we provided were just examples of >>things that one could mention? I have tried to make this clearer in >>the revised text. I agree with these three requirements. The revised text is more clear. >> >> > Details on how the specimens were stored until use may >>optionally be provided. For >> >> > example, if frozen samples were used, information should be >>provided regarding storage >> >> > temperature and duration of storage. >> >> > >> >> > I realize that this is optional. Nonetheless, this type of >>information is, in our experience, >> >> > never reported. Mentioning this optional requirement could >>reinforce a reviewer's >> >> > perception that MISFISHIE is overkill. >> >>We understand your point here and debated it here at length. We >>note that we have seen some papers where this information was >>revealed. The highlights of the debate may be summarized as: >> >>1. Is there a role for optional elements in MISFISHIE? Would it be >>better to remove anything optional and just have bare minimum? Or >>would it be advantageous to specify some optional "best practices" >>that we encourage authors to provide if they wish, but not lay it >>out as a requirement? Maybe I'll throw this point by itself out to >>the list to see what reactions are. >> >>2. We were inclined to leave it in, and see what our reviewers think. >> >>3. Larry suggested raising this point in the submission cover letter. >> >>Anyway, I have removed the mention of storage parameters. In general, I think it is a good idea to specify some optional best practises. >> >> > Reporters (probes or antibodies): >> >> > >> >> > Protocol(s) for how the reporters were designed and produced or >>the source from which >> >> > they were obtained. For GFP-like experiments[HC1], the promoter >>sequence should be specified. >> >> > >> >> > o For reporters purchased from a company, the company name and >> >> > catalogue number should be provided, as well as the web site, if >>available, >> >> > that provides details of the specifications. In addition, key >>aspects in the >> >> > specifications should be repeated since catalogue numbers and company >> >> > literature may not be available in the future. >> >> > >> >> > All these points are well taken. However, given that it is >>already a requirement to provide >> >> > the sequence of the promoter, this is too much additional detail >>as far as minimal >> >> > requirements are concerned. >> >>Okay, you're right, this could be quite burdensome. The promoter >>part will be moved elsewhere, it doesn't belong in this passage. >>We propose to remove "as well as the web site" and everything after >>that. Therefore, the new text would just require company names and >>catalog numbers when reporters are purchased from companies. Would >>you agree that this should be minimum, or do you think this is >>still too much? This is now the same as the requirement on your GXD >>gxd_submission_guidelines.shtml page (referenced at bottom). As far as commercial antibodies are concerned, I think it is a good idea to ask for company names and catalog numbers. For a nucleotide probe, providing the sequence of the probe should be sufficient. >> >> > Both positive and negative measurements of staining relevant to >>the experiment >> >> > should be reported. It is quite useful to provide negative >>expression results; it is >> >> > understood that a negative result is actually an upper limit to >>the expression level, >> >> > where the limit is usually not well known. >> >> > >> >> > Sounds good. >> >> > >> >> > For example: >> >> > Luminal epithelial cell: present >> >> > Basal epithelial cell: absent >> >> > etc. >> >> > or: >> >> > Luminal epithelial cell: 90% present, 10% equivocal, 0% absent >> >> > Basal epithelial cell: 0% present, 20% equivocal, 80% absent >> >> > etc. >> >> > >> >> > This detail in data annotation would be ideal and, obviously, it >>is just an example. >> >> > Nonetheless it is worth pointing out that it makes a huge >>difference if one has to annotate >> >> > a section of a part of a tissue or of a whole organism. The >>specification of the required >> >> > detail of annotations must take this into account, i.e. the >>amount of work must be kept >> >> > reasonable for the task at hand. Other issues to consider here >>are space limitations in >> >> > journal publications and readability of the manuscript. >>Electronic data submissions to >> >> > pertinent public databases have their place here but, again, the >>requirements must be >> >> > kept reasonable. >> >>Although we don't present it here in tabular form in the example, >>we'll clarify that it would best be presented in tabular form, >>either within the manuscript or as supplemental material depending >>on the scope of the data. >> >>Duncan further commented: >> >>I also still feel that we should be careful to distinguish the >>requirements for results from the interpretation and annotation of >>these results. In particular, I think it is unwise to require that >>people annotate all of their data. Many in situ experiments are >>targeted at specific organs, but incidentally give results for >>other tissues. A requirement for complete annotation runs the >>danger of pressing people to make inexpert annotations of the >>latter in order to push through publication. If a study targets a specific organ and does not annotate the results for other tissues, I think it would be very helpful to state this fact explicitly in the paper. >> >>I think we made it clear that we aren't requiring totally complete >>annotation of all possible visible structural units, but rather >>just of the structural units of interest to the paper, i.e. if >>positive expression for one reporter is listed for a structural >>unit, then that structural unit should be characterized for all >>other reporters, too. Yes, this is what I would like to see, too. This is particularly true for your example below: >> >>Do I correctly interpret your comments to reflect that I could say >>that "CD1 is present in cell type A, CD2 is present in cell type B, >>and CD3 is present in cell types B and C" and that by itself should >>be entirely sufficient as a minimum. i.e. I have not explicitly >>provided any information on whether CD1 is expressed in cell type B >>and that is fine. >> As previously written/envisioned we would require a table >>indicating calls for CD1,CD2,CD3 in cell types A,B,C, i.e. 9 >>measurements. But you feel this is too much? This is a critical >>point here. If we relax this requirement, then there's essentially >>no improvement of the status quo and negative results won't get >>reported. With some regret, I have added a sentence indicating >>that nulls are allowed. My opinion regarding your example is: If the objective of a study is to examine / compare the expression of CD1, CD2, and CD3 in three different cell types, information on whether CD1 is expressed in cell type B should be provided explicitly. >> >> > Protocol for the characterization and information about the >>basic technique for >> >> > characterizing the assays. For example, this information may >>include how many >> >> > observers performed the characterizations, assessment of >>inter-observer variability, >> >> > whether the characterizations were performed from the images >>themselves or visually >> >> > through instrument, any exceptions or assumptions made in >>characterizing the data, etc. >> >> > >> >> > This sounds to me like maximum rather than minimum information. >>Further, in most >> >> > cases the real number of observers is 1 (or maybe 2 max) and >>there is no assessment of >> >> > inter-observer variability. >> >> > >> >> > One general problem with the manuscript might be that it tries >>to describe a best practices >> >> > effort by the NIDDK consortium, a very valid goal in and of >>itself and, at the same time, >> >> > suggests these best practices as the minimum standard for the >>whole community. >> >> > Researchers doing these types of studies might be more receptive >>to the paper if it is >> >> > more pitched as a best practices effort. >> >>We propose to strike the inter-observer part completely, and >>rephrase the "protocol for characterization" as an optional >>best-practice, but not a requirement, as with some of the other >>protocols. Sounds good. Thanks also for your other comments and explanations. Martin |
From: Susanna S. <sa...@eb...> - 2005-07-13 14:01:38
|
John, no really, but please use our review: Sansone, S.A, Morrison, N., Rocca-Serra, P., Fostel, J. 2005.=20 Standardization initiatives in the (eco)toxicogenomics domain: a review.=20 Comp. Funct. Genomics. 8: 633-641. To cut a long story short, I am waiting to complete the loading of the=20 ILSI datasets into ArrayExpress. This will be the first large=20 toxicogenomics datasets to be publicly available. Then I will write a=20 paper on the tox infrastructure around ArrayExpress, covering both=20 MIAME/Tox (the specification) and Tox-MIAMExpress (the tool). See you in few weeks in DC. Ciao, Susanna John Quackenbush wrote: >Susanna, > >Is there a citable "official" MIAME-Tox publication yet? > >JQ > >Susanna Sansone wrote: > > =20 > >>Hi Eric, >> >>the "Summary recommendations for standardization and reporting of >>metabolic analyses" by the SMRS group is out: >>http://www.nature.com/nbt/journal/v23/n7/abs/nbt0705-833.html >> >>You can add this to the MIAME, MIAPE checklist, in the text and >>references. >> >>Best regards, >>Susanna >> >>Eric Deutsch wrote: >> >> =20 >> >>>Hi everyone, >>> >>>Thank you to all who provided feedback to previous drafts. I have >>>addressed he latest comments from Martin and Duncan (and a few minor >>>things from others) and merged everything into another draft, and >>>here is the result: >>> >>>http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc >>><http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc= > >>> >>>Please read it over and send back any comments to the email list: >>> >>>mge...@li... >>><mailto:mge...@li...> >>> >>>If you do not belong to this list, your may join or view the archives >>>or membership at >>> >>>http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ >>> >>>Please do send comments if there are any more or you haven=92t chimed >>>in yet as soon as you can. Please get all comments back to me on >>>Wednesday July 13 at the latest. If you have questions, please let me >>>know or post the questions to the mailing list. >>> >>>I encourage you to forward this message to additional people whom you >>>think should see the manuscript but may not be on these distribution >>>lists. Note that you can view the current members list at the >>>SourceForge URL above. I will also be sending this manuscript to a >>>few more members of the community not on these lists, seeking input >>>from them. If you would like to suggest additional people and would >>>rather have me send them an email, let me know. >>> >>>Regards, >>> >>>Eric >>> >>> =20 >>> > > >------------------------------------------------------- >This SF.Net email is sponsored by the 'Do More With Dual!' webinar happe= ning >July 14 at 8am PDT/11am EDT. We invite you to explore the latest in dual >core and dual graphics technology at this free one hour event hosted by = HP, >AMD, and NVIDIA. To register visit http://www.hp.com/go/dualwebinar >_______________________________________________ >Mged-misfishie mailing list >Mge...@li... >https://lists.sourceforge.net/lists/listinfo/mged-misfishie > > =20 > --=20 Susanna-Assunta Sansone, PhD Nutri/Toxicogenomics Project Coordinator, EBI =09 The European Bioinformatics Institute email: sa...@eb... EMBL Outstation - Hinxton direct: +44 (0)1223 494 691 Wellcome Trust Genome Campus fax: +44 (0)1223 494 468 Cambridge CB10 1SD, UK =20 Project page: www.ebi.ac.uk/microarray/Projects/tox-nutri/index.html = =20 =20 |
From: John Q. <jo...@ji...> - 2005-07-13 13:40:27
|
Susanna, Is there a citable "official" MIAME-Tox publication yet? JQ Susanna Sansone wrote: > Hi Eric, > > the "Summary recommendations for standardization and reporting of > metabolic analyses" by the SMRS group is out: > http://www.nature.com/nbt/journal/v23/n7/abs/nbt0705-833.html > > You can add this to the MIAME, MIAPE checklist, in the text and > references. > > Best regards, > Susanna > > Eric Deutsch wrote: > >> Hi everyone, >> >> Thank you to all who provided feedback to previous drafts. I have >> addressed he latest comments from Martin and Duncan (and a few minor >> things from others) and merged everything into another draft, and >> here is the result: >> >> http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc >> <http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc= > >> >> Please read it over and send back any comments to the email list: >> >> mge...@li... >> <mailto:mge...@li...> >> >> If you do not belong to this list, your may join or view the archives >> or membership at >> >> http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ >> >> Please do send comments if there are any more or you haven=92t chimed >> in yet as soon as you can. Please get all comments back to me on >> Wednesday July 13 at the latest. If you have questions, please let me >> know or post the questions to the mailing list. >> >> I encourage you to forward this message to additional people whom you >> think should see the manuscript but may not be on these distribution >> lists. Note that you can view the current members list at the >> SourceForge URL above. I will also be sending this manuscript to a >> few more members of the community not on these lists, seeking input >> from them. If you would like to suggest additional people and would >> rather have me send them an email, let me know. >> >> Regards, >> >> Eric >> > |
From: Susanna S. <sa...@eb...> - 2005-07-13 13:00:04
|
Hi Eric, the "Summary recommendations for standardization and reporting of=20 metabolic analyses" by the SMRS group is out: http://www.nature.com/nbt/journal/v23/n7/abs/nbt0705-833.html You can add this to the MIAME, MIAPE checklist, in the text and reference= s. Best regards, Susanna Eric Deutsch wrote: > Hi everyone, > > Thank you to all who provided feedback to previous drafts. I have=20 > addressed he latest comments from Martin and Duncan (and a few minor=20 > things from others) and merged everything into another draft, and here=20 > is the result: > > http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc=20 > <http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc> > > Please read it over and send back any comments to the email list: > > mge...@li...=20 > <mailto:mge...@li...> > > If you do not belong to this list, your may join or view the archives=20 > or membership at > > http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ > > Please do send comments if there are any more or you haven=92t chimed i= n=20 > yet as soon as you can. Please get all comments back to me on=20 > Wednesday July 13 at the latest. If you have questions, please let me=20 > know or post the questions to the mailing list. > > I encourage you to forward this message to additional people whom you=20 > think should see the manuscript but may not be on these distribution=20 > lists. Note that you can view the current members list at the=20 > SourceForge URL above. I will also be sending this manuscript to a few=20 > more members of the community not on these lists, seeking input from=20 > them. If you would like to suggest additional people and would rather=20 > have me send them an email, let me know. > > Regards, > > Eric > --=20 Susanna-Assunta Sansone, PhD Nutri/Toxicogenomics Project Coordinator, EBI =09 The European Bioinformatics Institute email: sa...@eb... EMBL Outstation - Hinxton direct: +44 (0)1223 494 691 Wellcome Trust Genome Campus fax: +44 (0)1223 494 468 Cambridge CB10 1SD, UK =20 Project page: www.ebi.ac.uk/microarray/Projects/tox-nutri/index.html = =20 =20 |
From: G. S. B. <gb...@jh...> - 2005-07-06 21:58:36
|
Dear Eric, Larry, Alvin, Christian, John and others on the MISFISHIE list that I don't know. Thanks for working on this, it could lead to badly needed improvement in published immunostain, FISH, and related data. I joined the list only recently and have a few short comments that I can elaborate on in another email (I'll be back from vacation next week) if you find any of them helpful. These are in no particular order: -In our own tissue microarray/antibody staining work, the thing we find most difficult using the literature right now in attempts to validate/replicate studies are the lack of specific replicable positive and negative control data, and the lack of specificity on the exact antibody clone used. For example, there are several papers studying MDM2 expression in prostate cancer, but none report what they used as a positive or negative control (there is a lung cancer cell line commonly used for this), and in addition, several suppliers provide more than one MDM2 antibody, so while all publications include the name of the supplier, a fair number don't specify the clone, making truly impossible to replicate without getting in touch with the authors, who often don't remember. I think the checklist would be a good place to prompt authors/reviewers to remember to include the supplier_clone_name. The specific positive and negative controls used (and how prepared) should also be specifically requested...and I think it would be good to encourage people to use controls that are freely available...otherwise the chain leading to replication or refutation will be tenuous. -For each tissue studied, the anatomic source beyond species should be specified...ie, the liver, adrenal, lung, right lower lobe of lung, etc. I looked through the specification several times, I must be missing this, but I couldn't find where the specification asks for this. -Antibodies have Class/Subclass and Type/Subtypes categorizations. They aren't that complex and could be defined specifically in the checklist or as a reference table to the checklist to help the author/reviewer comply. I think the term "isotype" used currently in the specfication refers to Class, but I'm not sure. -There may be a bit of a conflict related to images in that those who are reporting results from reading images directly through a microscope or from a live video image, if they comply, could easily provide images that are of regions that were not the same as those used in their scoring. Based on the following study that we did a while ago: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=11331959&query_hl=2 I think there is strong enough evidence to encourage those publishing immunostaining or FISH results to always define and store the set of images to be scored first, then score the electronically stored image.... in this paper, we were struck by the many advantages (in terms of replication, decreased intraobserver and interobserver variability, etc) of doing these studies this way, and only this way. You have an undefined kettle of fish otherwise. -Another point I think that would be worth making in the paper is that there is still a lot of room for research in defining scientific best practices for doing and reporting these types of studies....it might be worth adding this to your statement that the group will be working to improve the standard over time. Let me know if any of this is helpful, the small public library I'm in in Blue Hill Maine is closing, so I have to stop here. All the best, Steve Bova G. Steven Bova M.D. Asst Prof. of Pathology Asst Prof of Health Sci Informatics, Genetic Medicine, Oncology, Urology Department of Pathology Johns Hopkins University School of Medicine Carnegie Building Rm 628 600 North Wolfe St. Baltimore, MD 21287-6417 Email: gb...@jh... Fax: 410-614-7620 Voicemail: 410-614-5957 |
From: John Q. <jo...@ji...> - 2005-07-05 21:21:11
|
Eric, The 0.02ms review - my address is wrong. It should be: Dana-Farber Cancer Institute, 44 Binney Street, M232, Boston, MA 02115 Eric Deutsch wrote: > Hi everyone, > > Thank you to all who provided feedback to previous drafts. I have > addressed he latest comments from Martin and Duncan (and a few minor > things from others) and merged everything into another draft, and here > is the result: > > _http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v1__5__.doc_ > <http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc> > > Please read it over and send back any comments to the email list: > > _mg...@li..._ > <mailto:mge...@li...> > > If you do not belong to this list, your may join or view the archives > or membership at > > _http://lists.sourceforge.net/lists/listinfo/mged-misfishie/_ > > Please do send comments if there are any more or you haven't chimed in > yet as soon as you can. Please get all comments back to me on > Wednesday July 13 at the latest. If you have questions, please let me > know or post the questions to the mailing list. > > I encourage you to forward this message to additional people whom you > think should see the manuscript but may not be on these distribution > lists. Note that you can view the current members list at the > SourceForge URL above. I will also be sending this manuscript to a few > more members of the community not on these lists, seeking input from > them. If you would like to suggest additional people and would rather > have me send them an email, let me know. > > Regards, > > Eric > |
From: Eric D. <ede...@sy...> - 2005-07-05 21:20:41
|
Hi everyone, one of the items that came up in the previous long email was the issue of optional elements. For example, the current draft says: Optionally as a best practice, the protocol for the characterization and information about the basic technique for characterizing the assays. I would like to pose to all of you, do you think it's a good idea to include (a very few) optional items? These specifically would be items where we would like to suggest that furnishing the information in question is a very good idea, but we won't make it a strict requirement. Would this promote better practices in reporting information? Or would this just cause confusion and a more negative reaction? Let us know what you think! Thanks, Eric |
From: Eric D. <ede...@sy...> - 2005-07-05 21:09:38
|
Hi everyone, Thank you to all who provided feedback to previous drafts. I have addressed he latest comments from Martin and Duncan (and a few minor things from others) and merged everything into another draft, and here is the result: http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v15.doc Please read it over and send back any comments to the email list: mge...@li... If you do not belong to this list, your may join or view the archives or membership at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ Please do send comments if there are any more or you haven't chimed in yet as soon as you can. Please get all comments back to me on Wednesday July 13 at the latest. If you have questions, please let me know or post the questions to the mailing list. I encourage you to forward this message to additional people whom you think should see the manuscript but may not be on these distribution lists. Note that you can view the current members list at the SourceForge URL above. I will also be sending this manuscript to a few more members of the community not on these lists, seeking input from them. If you would like to suggest additional people and would rather have me send them an email, let me know. Regards, Eric |
From: Eric D. <ede...@sy...> - 2005-07-05 21:03:45
|
> From: mge...@li... [mailto:mged-misfishie- > ad...@li...] On Behalf Of Martin Ringwald >=20 > Following Eric's suggestion, I am posting an updated and extended > version of my previous comments to the whole list - see attached pdf > file. I hope this file format works for everyone. Hi Martin and Duncan, thank you for posting your comments on the manuscript here. I spoke at length with Larry about how to address your concerns. We have made changes in response to nearly all of them, and would be grateful if you would reread the latest draft and let us know whether you are happy with the revision, or if you have further comments. Although there have been no other responses to your comments, I highly encourage others out there to add any comments they wish. > Following Eric's suggestion, I am posting an updated version of my previous comments > to the whole list. My comments referring to MISFISHIE_v9.2.doc were originally sent, > on April 2, to Eric and several others who were listed as authors at the time. Since then, > Eric posted excerpts of my letter to the whole list but, unfortunately, the excerpts didn't > convey the whole picture. This posting refers to MISFISHIE_v14.doc. The comments are > based on our work on the Mouse Gene Expression Database (GXD) project that has > extensive experience with RNA in situ hybridization and immunohistochemistry data and > with the extraction and annotation of these types of data from the literature. > > I think it is a good idea to develop and promote a minimal information standard for RNA > in situ hybridization and immunohistochemistry experiments following the example of > MIAME, and I am more than happy to contribute to this effort. However, I find myself in > a dilemma with the current version of the manuscript. I agree with the big picture put > forward in Figure 1 and with many other basic points of the manuscript as well as with > the statement that the community, and databases such as GXD and EMAGE, would > benefit from more detailed experimental descriptions. However, in my opinion, much of > the information requested goes beyond what I would consider to be minimal > requirements. This is certainly a fair opinion. If you think it is too demanding, then it almost certainly is. I hope to work with you to strike the right balance of improving on the status quo without causing a negative reaction to the overall goal. > Minimal requirements would be applied to the whole community including those > researchers pursuing large-scale screens and those who run conventional laboratories, as > well as researchers from many different disciplines. These groups are approaching their > experiments in different ways, and they will have different challenges in complying with > the standard. For example, for researchers generating RNA in situ hybridization data on a > large-scale under unified conditions, it will be relatively straightforward to record the > nitty-gritty details of experimental parameters, but they will find it more challenging to > annotate all the results in adequate detail. On the other hand, small-scale laboratories, > who continually refine their experimental conditions, often on an individual (e.g. > postdoc) basis, will find it more difficult to report all experimental parameters, but they > are in a better position to evaluate and annotate their results. Also, there is a huge > difference between having to annotate a section from a relatively simple tissue and one > that shows a section of the whole embryo. Finally, I should point out that RNA in situ > hybridization and immunohistochemistry, while still evolving, are long established > techniques with a long history of the way in which such data are published - for better or > worse: for better, because the assays and their limitations are well understood and people > often use standard protocols; for worse because pertinent material and methods sections > become more cryptic with time. Yes, we should try to aim for a more complete > description of the experiments; we should not aim for the lowest common denominator. > However, I believe we should aim for a minimal set of requirements that are, from a > scientific point of view, sensible for the different sectors of the community, including > databases. I agree. > MIAME was developed in the hope that microarray data from different laboratories could > eventually be reproduced and, at least to some extent, compared at the quantitative level. > Therefore, it tries to capture many details about all the steps in a microarray experiment > that could possibly influence the expression levels detected. RNA in situ hybridization > and immunohistochemistry data are different from microarray data in that there is a much > more limited expectation of quantitative comparability. At last year's EMAGE Advisory > Board Meeting, I asked leading experts in RNA in situ hybridization technology, such as > David Wilkinson who wrote a book about the subject, if there is any development that > could lead to quantitative comparability across laboratories. The consensus at that time > was there is no such development in sight. I think we have to take this situation into > account when we discuss the minimum requirements. There is no point in requiring > details that, ultimately, will be of limited use but put undue burden on researchers. We > should keep the barriers reasonably low but high enough to ensure that experimental > descriptions include the details that are really needed. Fair enough. Several rounds ago, before recent reduction of some requirements, there were many requirements that were included because we thought they could possibly influence whether expression was detected. But after you pointed them out, we realized that several of these requirements might be unduly burdensome to some types of experimental descriptions, and we dropped them. > Here are a few specific examples of where I think the manuscript requests too much in > terms of minimal information: > > Experimental Design: > > Total number of hybridizations/stains performed in the experiment: a hybridization/stain > is defined as an assay of a single specimen with a single reporter. Thus, the result of a > tissue microarray consisting of a 10 x 10 array of different tissues would be counted as > 100 immunostains. If replicates or reruns are a component of the experimental design, > provide details that should include number of replicates per tissue, per reporter, etc. > > This might make sense for tissue arrays but I don't think it can be readily applied to > conventional RNA in situ and immunohistochemistry studies (as a minimal requirement). > Of course, people make replicate experiments but most of the time they don't do this in a > systematic fashion and regular papers rarely, if ever, provide this type of information. > (Yes, you could argue that in these cases replicates are not a component of the > experimental design.) We didn't understand why you don't think these two requirements: - total # of stains reported on - *IF* there are replicates as part of the of the design, tell us are applicable to conventional ISH and IHC. We did find several papers that mentioned replicate information ("each antibody assayed with 3 - 5 separate tissues"). If the papers were silent on replicates, we assumed there were none. Would it be a bad thing to force authors to state or imply "we performed no replicates"? Duncan commented: I think authors should be required to describe the number of replicates in their data. As Martin says, people often don't do a strict series of replicates, but I would say that they should be required to show that they have done enough replicates to substantiate the results. MISFISHIE currently doesn't even require that they show that they have done enough replicates. It only requires that they reveal what they did. It would be up to a reviewer (or the reader) to decide if it was "enough". In our view, this requirement should provoke authors to write 1 sentence in their manuscript saying what the total assay count was and how replicates are done if they are done at all. This does not seem burdensome. If they don't do replicates, that's fine as far as MISFISHIE goes. If the reviewer is happy about no replicates, then all is well. We don't feel strongly about this point. We'd like to ask you what you suggest we write here? Strike it completely? Reword? Please suggest what you would like to see in the manuscript regarding this. > Biomaterials and Treatments: > > Manner of preparation of the specimens for the study. Information required includes the > nature of the samples: e.g. whole tissue, tissue sections, thickness of sections, whole > cells, > or sections of cells; manner in which the specimens were prepared for the experiments, > e.g. fixation with type of fixative and duration of fixation vs. fresh, non-fixed, non-frozen > specimens or frozen specimens, sections mounted on slides versus sections floating in > reagents, nature of the slides on which sections were mounted; and the protocols used. > > It would be good to have this data but I don't think information such as the nature of the > slides on which sections were mounted should be mandatory or is even that useful. We will strike the "nature of the slides". We will also adjust the punctuation a little to emphasize that there are only 3 requirements intended here: 1. nature of the samples 2. manner in which the specimens were prepared for the experiments 3. protocols used Everything else in the paragraph is examples of the sorts of things that should be considered for inclusion. We are not intending to require all of the examples provided, but rather only require that the paper describe the three above points so that the experiment can be properly assessed and replicated (to the satisfaction of the reviewer). Do you disagree with any of these three requirements? Or was it just unclear that the examples we provided were just examples of things that one could mention? I have tried to make this clearer in the revised text. > Details on how the specimens were stored until use may optionally be provided. For > example, if frozen samples were used, information should be provided regarding storage > temperature and duration of storage. > > I realize that this is optional. Nonetheless, this type of information is, in our experience, > never reported. Mentioning this optional requirement could reinforce a reviewer's > perception that MISFISHIE is overkill. We understand your point here and debated it here at length. We note that we have seen some papers where this information was revealed. The highlights of the debate may be summarized as: 1. Is there a role for optional elements in MISFISHIE? Would it be better to remove anything optional and just have bare minimum? Or would it be advantageous to specify some optional "best practices" that we encourage authors to provide if they wish, but not lay it out as a requirement? Maybe I'll throw this point by itself out to the list to see what reactions are. 2. We were inclined to leave it in, and see what our reviewers think. 3. Larry suggested raising this point in the submission cover letter. Anyway, I have removed the mention of storage parameters. > Reporters (probes or antibodies): > > Protocol(s) for how the reporters were designed and produced or the source from which > they were obtained. For GFP-like experiments[HC1], the promoter sequence should be specified. > > o For reporters purchased from a company, the company name and > catalogue number should be provided, as well as the web site, if available, > that provides details of the specifications. In addition, key aspects in the > specifications should be repeated since catalogue numbers and company > literature may not be available in the future. > > All these points are well taken. However, given that it is already a requirement to provide > the sequence of the promoter, this is too much additional detail as far as minimal > requirements are concerned. Okay, you're right, this could be quite burdensome. The promoter part will be moved elsewhere, it doesn't belong in this passage. We propose to remove "as well as the web site" and everything after that. Therefore, the new text would just require company names and catalog numbers when reporters are purchased from companies. Would you agree that this should be minimum, or do you think this is still too much? This is now the same as the requirement on your GXD gxd_submission_guidelines.shtml page (referenced at bottom). > * Protocol used to produce the hybridization or immunostain. This should include a > description of how the section was mounted onto the slide/substrate and > treatments of the section, e.g. immunohistochemistry protocol, inclusive of > parameters such as buffer, temperature, post-wash conditions, etc. Also include: > > o What steps, if any, were taken to decrease non-specific reaction product. > For example in immunoperoxidase experiments there might be preincubation > of the specimen preparation with (a) an albumin solution to > block non-specific binding of protein, (b) a peroxide solution to block > signal due to endogenous peroxidase. > > Yes, this information is important. However, because many researchers use well established > protocols, currently much of this information is included in papers via > reference to pertinent publications. MISFISHIE should provide this option as well. We agree. We will add here the sentence already in a different protocol section saying that protocols already published may be referenced if followed exactly. > Imaging data: > > The images should represent the range of gene expression at different magnifications (or > spatial resolutions). > > I see the good intention but the scope is ill defined here and it could potentially be very > large and well beyond what authors and journals are willing to accept. You're right. We'll drop this. > Image acquisition protocol (part of Image acquisition parameters) > > This information is much more relevant for microarray data - as far as minimum > requirements are concerned. The example provided in Additional File 1 states that a > specific type of camera and photoshop was used but this doesn't tell you anything. For > most cases, it doesn't matter what brand of camera you use, as long as you have a decent > one, and optimal data acquisition parameters are highly dependent on local conditions. Okay, fair enough. We will drop this, too. > Image Characterizations: > > * Intensity scale, ideally choosing one from the MGED Ontology. For example, a > three-level scale of present, absent, or equivocal might be appropriate for > evaluating immunohistochemistry stains. However, any scale that the > investigators feel is appropriate may be used as long as each gradation of > intensity in the scale is defined in a manner that enables an independent > investigator to understand or apply the same criteria. > > * Per each structural unit (relevant to the experiment) in each assay (or in each > image), quantitative measurement or an estimate of: > > o Staining intensity, or the fraction of the structural unit's population > exhibiting each intensity (see example below) > > I agree with what is said about the intensity scale. However, in most cases quantitative > measurements will be very difficult if not impossible to provide. The intensity scale is specified by the authors. It may be as simple as present/absent. No quantitative measurements beyond that are required. I've removed the reference to quantitative measurement. > Both positive and negative measurements of staining relevant to the experiment > should be reported. It is quite useful to provide negative expression results; it is > understood that a negative result is actually an upper limit to the expression level, > where the limit is usually not well known. > > Sounds good. > > For example: > Luminal epithelial cell: present > Basal epithelial cell: absent > etc. > or: > Luminal epithelial cell: 90% present, 10% equivocal, 0% absent > Basal epithelial cell: 0% present, 20% equivocal, 80% absent > etc. > > This detail in data annotation would be ideal and, obviously, it is just an example. > Nonetheless it is worth pointing out that it makes a huge difference if one has to annotate > a section of a part of a tissue or of a whole organism. The specification of the required > detail of annotations must take this into account, i.e. the amount of work must be kept > reasonable for the task at hand. Other issues to consider here are space limitations in > journal publications and readability of the manuscript. Electronic data submissions to > pertinent public databases have their place here but, again, the requirements must be > kept reasonable. Although we don't present it here in tabular form in the example, we'll clarify that it would best be presented in tabular form, either within the manuscript or as supplemental material depending on the scope of the data. Duncan further commented: I also still feel that we should be careful to distinguish the requirements for results from the interpretation and annotation of these results. In particular, I think it is unwise to require that people annotate all of their data. Many in situ experiments are targeted at specific organs, but incidentally give results for other tissues. A requirement for complete annotation runs the danger of pressing people to make inexpert annotations of the latter in order to push through publication. I think we made it clear that we aren't requiring totally complete annotation of all possible visible structural units, but rather just of the structural units of interest to the paper, i.e. if positive expression for one reporter is listed for a structural unit, then that structural unit should be characterized for all other reporters, too. I am inferring that this is still perceived as too burdensome? Do I correctly interpret your comments to reflect that I could say that "CD1 is present in cell type A, CD2 is present in cell type B, and CD3 is present in cell types B and C" and that by itself should be entirely sufficient as a minimum. i.e. I have not explicitly provided any information on whether CD1 is expressed in cell type B and that is fine. As previously written/envisioned we would require a table indicating calls for CD1,CD2,CD3 in cell types A,B,C, i.e. 9 measurements. But you feel this is too much? This is a critical point here. If we relax this requirement, then there's essentially no improvement of the status quo and negative results won't get reported. With some regret, I have added a sentence indicating that nulls are allowed. > Protocol for the characterization and information about the basic technique for > characterizing the assays. For example, this information may include how many > observers performed the characterizations, assessment of inter-observer variability, > whether the characterizations were performed from the images themselves or visually > through instrument, any exceptions or assumptions made in characterizing the data, etc. > > This sounds to me like maximum rather than minimum information. Further, in most > cases the real number of observers is 1 (or maybe 2 max) and there is no assessment of > inter-observer variability. > > One general problem with the manuscript might be that it tries to describe a best practices > effort by the NIDDK consortium, a very valid goal in and of itself and, at the same time, > suggests these best practices as the minimum standard for the whole community. > Researchers doing these types of studies might be more receptive to the paper if it is > more pitched as a best practices effort. We propose to strike the inter-observer part completely, and rephrase the "protocol for characterization" as an optional best-practice, but not a requirement, as with some of the other protocols. > Additional comments: > > The review considered that more than 90% of the papers were compliant with > MISFISHIE sections 1 and 2 (Experimental Design; Biomaterials and Treatments). > Compliance for sections 3 and 4 (Reporters and Staining) was about 75%. Section 5 > (Imaging Data) proved to be the most troublesome, with only 16% of the articles > compliant. Finally, about 47% complied with section 6 (Image Characterizations) > > If we take the requirements put forward in the main part of the manuscript literally, then > these numbers are significantly higher than we would have expected from our own > literature curation work. GXD's focus is on endogenous gene expression during mouse > development. At this point, we have indexed nearly 11,000 references. We have coded > over 19,000 assays in detail, including about 7,700 RNA in situ hybridization and 1,250 > immunohistochemistry experiments (the other types of assays are RT-PCR, Northern > blot, Western blot and RNAse and Nuclease 1 protection data). To give just two > examples: in our experience, except for RT-PCR, we rarely, if ever, see papers that > describe how many replicates the authors did (section 1); and the number of papers that As discussed at the top, we interpreted the standard such that replicates need only be mentioned if they are part of the experimental design. If it is not mentioned, it was assumed to mean no replicates, which is fine. So lack of number of replicates does not lead to non-compliance. > specify the strain or provide all the details regarding the preparation of the specimens > (section 2) is significantly lower than 90%. Strain itself is not listed as an explicit requirement, but rather merely provided in a list of attributes that might be relevant. It would ultimately be up to the reviewer to decide whether missing strain information would constitute insufficient specimen attributes to properly assess or reproduce the study. > Strain information is definitively important for mouse studies yet we find that mouse > developmental biologists pay less attention to strain information than mouse geneticists. > This illustrates that even closely related communities can have different de facto > standards - and it illustrates the importance of common standards such as MISFISHIE. > On the other hand, strain information is, of course, not relevant for human and many > other species. And for the latter reason, MISFISHIE does not explicitly require strain. It is merely one of the possible attributes of each specimen that should be provided if it is important to properly assess or reproduce the study (in the opinion of the reviewer, of course). > Anyway, the discrepancy between the numbers of your literature survey and our own > experience could have several reasons: > > 1. Mouse developmental biologists do not meet the general standard when it comes to the > description of their data in publications. I do not think that this is the case. I have insufficient experience to judge, but I suspect you're right. However, we probably did not fail any papers because strain specifically was not provided, whereas it sounds like you would have. > 2. The compliance criteria are softer than the manuscript suggests. Given that scores of 8 > and 9 were still considered passing and that each final section score was a subjective > weighted aggregate from the individual subsections it is not entirely clear what, exactly, > compliance means. Perhaps we have been too vague. I will tighten up the language here. 8 was generally considered a low pass, 9 a medium pass, and 10 a high pass. We did this for our own internal data collection to try to provide reviewers with more flexibility than just pass or fail, realizing that the reviews are somewhat subjective. > 3. The 32 articles chosen for review are not a representative sample. I realize that it is > very difficult to come up with a representative sample. However, the selection of the > journals can have an impact - often times, high profile journals have stronger space > limitations. Following the idea that a paper should provide enough detail to allow others > to interpret and reproduce experiments, I think it would be a good idea to provide > citations for the 32 surveyed publications in the MISFISHIE paper. Okay, we will provide citations for the 32 papers. It is true that our sample was mostly human-specimen papers from higher-profile or pathology journals. There is not a big overlap with the papers you see in GXD. > More importantly, a general-purpose repository to which researchers could submit their > images for permanent storage with accession numbers for publications would be very > valuable for facilitating MISFISHIE compliance and in realizing the full value of these > data for future research. BioImage is such a repository already under construction at > http://www.bioimage.org/. > It may therefore not be appropriate to enforce section 5 of MISFISHIE until a suitable > image repository is available, and a consensus in the research community that requiring > images is appropriate, acceptable and achievable is established. > > The paragraphs above only mention the general BioImage database that is still under > construction but they don't mention, in this context, other public databases that already > deal with these types of data, including image data, such as FlyView, ZFIN (zebrafish), > Xenbase, GXD and EMAGE. (GXD currently holds almost 19,000 images linked to > detailed annotations). A central image repository might be helpful in particular for > scientific domains for which no appropriate databases exist. However, the real bottleneck > is the proper annotation and integration of the image data. This requires a lot of effort and > domain-specific expertise. Even if a paper would fulfill all the requirements put forward > in the current version of MISFISHIE, it is very likely that the data will still need > additional curatorial work. For me it is hard to imagine that this work can be done by a > general image repository. You're right, of course. I will include references to these databases, and point out the lack of repository for human studies and for many other organisms. But I think it is true that until there is a repository that will accept images from any organism, we have a problem with requiring images for compliance. I do not believe that there needs to be a lot of curatorial work to satisfy MISFISHIE. All that would be required is that images be stored in some public repository with some shared tag that would allow a reader of a paper to find the images behind a certain expression call. There is no requirement in MISFISHIE that the expression calls themselves be in the database. The only requirement is that the images are available and tagged minimally with specimen and reporter. > Ideally, a paper specifying the minimum information for in situ hybridization and > immunohistochemistry experiments would represent the consensus, or at least the input, > of a good number of researchers who deal with this type of data, including > experimentalists and those of us who maintain pertinent databases. Therefore, I think it > would be good to receive more feedback from people who publish or store RNA in situ or > immunohistochemistry data. I agree. After this round of changes, I will start circulating more widely. > I hope these comments are helpful, They are very helpful. Thank you for all the time and effort you have put into examining the manuscript and laying down your concerns in writing. I would grateful if you would respond to the questions herein, if we have resolved your concerns or have still not gone far enough. Thank you, Eric > Best wishes, > Martin > > PS: FYI, our current guidelines for electronic submission of expression data, including > RNA in situ and immunohistochemistry data are accessible at > http://www.informatics.jax.org/mgihome/GXD/GEN/gxd_submission_guidelines .shtml. >=20 >=20 >=20 |
From: Duncan D. <du...@hg...> - 2005-06-17 08:39:38
|
Dear Eric, Thank you. Again, I appreciate the effort you are making on this. Best wishes, Duncan Eric Deutsch wrote: > Dear Martin and Duncan, thank you very much for your feedback. I met > the Larry the other day and we went through all of Martin's comments in > detail. I will soon be making changes to the manuscript and address > each of the points you both have made. If anyone else wishes to commen= t > on these or additional points, please post them here at any time. >=20 > Thank you! > Eric >=20 >=20 >=20 >>-----Original Message----- >>From: mge...@li... >=20 > [mailto:mged-misfishie- >=20 >>ad...@li...] On Behalf Of Duncan Davidson >>Sent: Thursday, June 16, 2005 10:39 AM >>To: mge...@li... >>Subject: Re: [Mged-misfishie] comments on MISFISHIE >> >>Dear Eric, >> >>I agree with Martin Ringwald's comments and I think he has dealt very >>well with all the points that I too have had reservations about. The >>only exception is that I think authors should be required to describe >>the number of replicates in their data. As Martin says, people often >>don't do a strict series of replicates, but I would say that they >=20 > should >=20 >>be required to show that they have done enough replicates to >>substantiate the results. I also still feel that we should be careful >=20 > to >=20 >>distinguish the requirements for results from the interpretation and >>annotation of these results. In particular, I think is is unwise to >>require that people annotate all of their data. Many in situ >=20 > experiments >=20 >>are targeted at specific organs, but incidentally give results for >=20 > other >=20 >>tissues. A requirement for complete annotation runs the danger of >>pressing people to make inexpert annotations of the latter in order to >>push through publication. >> >>Like Martin, I strongly support your efforts to develop a minimum >>standard for in situ data. But I also support Martin's suggestions >=20 > that >=20 >>you can get more comment from people who have done a lot of in situs - >>both hypothesis-driven and the large screens in different species. >> >>Again, thanks for your efforts. >> >>Best wishes, >> >>Duncan >> >> >> >> >> >>Martin Ringwald wrote: >> >> >>>Following Eric's suggestion, I am posting an updated and extended >>>version of my previous comments to the whole list - see attached pdf >>>file. I hope this file format works for everyone. >>> >>>Best wishes, >>> >>>Martin >> >> >> >>------------------------------------------------------- >>SF.Net email is sponsored by: Discover Easy Linux Migration Strategies >>from IBM. Find simple to follow Roadmaps, straightforward articles, >>informative Webcasts and more! Get everything you need to get up to >>speed, fast. http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dcl= ick >>_______________________________________________ >>Mged-misfishie mailing list >>Mge...@li... >>https://lists.sourceforge.net/lists/listinfo/mged-misfishie >=20 >=20 >=20 >=20 >=20 > ------------------------------------------------------- > SF.Net email is sponsored by: Discover Easy Linux Migration Strategies > from IBM. Find simple to follow Roadmaps, straightforward articles, > informative Webcasts and more! Get everything you need to get up to > speed, fast. http://ads.osdn.com/?ad_idt77&alloc_id=16492&op=CCk > _______________________________________________ > Mged-misfishie mailing list > Mge...@li... > https://lists.sourceforge.net/lists/listinfo/mged-misfishie |
From: Eric D. <ede...@sy...> - 2005-06-17 05:03:04
|
Dear Martin and Duncan, thank you very much for your feedback. I met the Larry the other day and we went through all of Martin's comments in detail. I will soon be making changes to the manuscript and address each of the points you both have made. If anyone else wishes to comment on these or additional points, please post them here at any time. Thank you! Eric > -----Original Message----- > From: mge...@li... [mailto:mged-misfishie- > ad...@li...] On Behalf Of Duncan Davidson > Sent: Thursday, June 16, 2005 10:39 AM > To: mge...@li... > Subject: Re: [Mged-misfishie] comments on MISFISHIE >=20 > Dear Eric, >=20 > I agree with Martin Ringwald's comments and I think he has dealt very > well with all the points that I too have had reservations about. The > only exception is that I think authors should be required to describe > the number of replicates in their data. As Martin says, people often > don't do a strict series of replicates, but I would say that they should > be required to show that they have done enough replicates to > substantiate the results. I also still feel that we should be careful to > distinguish the requirements for results from the interpretation and > annotation of these results. In particular, I think is is unwise to > require that people annotate all of their data. Many in situ experiments > are targeted at specific organs, but incidentally give results for other > tissues. A requirement for complete annotation runs the danger of > pressing people to make inexpert annotations of the latter in order to > push through publication. >=20 > Like Martin, I strongly support your efforts to develop a minimum > standard for in situ data. But I also support Martin's suggestions that > you can get more comment from people who have done a lot of in situs - > both hypothesis-driven and the large screens in different species. >=20 > Again, thanks for your efforts. >=20 > Best wishes, >=20 > Duncan >=20 >=20 >=20 >=20 >=20 > Martin Ringwald wrote: >=20 > > Following Eric's suggestion, I am posting an updated and extended > > version of my previous comments to the whole list - see attached pdf > > file. I hope this file format works for everyone. > > > > Best wishes, > > > > Martin >=20 >=20 >=20 > ------------------------------------------------------- > SF.Net email is sponsored by: Discover Easy Linux Migration Strategies > from IBM. Find simple to follow Roadmaps, straightforward articles, > informative Webcasts and more! Get everything you need to get up to > speed, fast. = http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclick > _______________________________________________ > Mged-misfishie mailing list > Mge...@li... > https://lists.sourceforge.net/lists/listinfo/mged-misfishie |
From: Duncan D. <du...@hg...> - 2005-06-16 17:39:23
|
Dear Eric, I agree with Martin Ringwald's comments and I think he has dealt very well with all the points that I too have had reservations about. The only exception is that I think authors should be required to describe the number of replicates in their data. As Martin says, people often don't do a strict series of replicates, but I would say that they should be required to show that they have done enough replicates to substantiate the results. I also still feel that we should be careful to distinguish the requirements for results from the interpretation and annotation of these results. In particular, I think is is unwise to require that people annotate all of their data. Many in situ experiments are targeted at specific organs, but incidentally give results for other tissues. A requirement for complete annotation runs the danger of pressing people to make inexpert annotations of the latter in order to push through publication. Like Martin, I strongly support your efforts to develop a minimum standard for in situ data. But I also support Martin's suggestions that you can get more comment from people who have done a lot of in situs - both hypothesis-driven and the large screens in different species. Again, thanks for your efforts. Best wishes, Duncan Martin Ringwald wrote: > Following Eric's suggestion, I am posting an updated and extended > version of my previous comments to the whole list - see attached pdf > file. I hope this file format works for everyone. > > Best wishes, > > Martin |
From: Martin R. <rin...@in...> - 2005-06-10 17:17:00
|
Following Eric's suggestion, I am posting an updated and extended version of my previous comments to the whole list - see attached pdf file. I hope this file format works for everyone. Best wishes, Martin |
From: Alvin L. <aliu@u.washington.edu> - 2005-05-18 15:57:23
|
MISFISHIE draft v14: Please respond latest May 23functional in = references. ----- Original Message -----=20 From: Eric Deutsch=20 To: mge...@li...=20 Sent: Tuesday, May 17, 2005 10:29 PM Subject: RE: [Mged-misfishie] MISFISHIE draft v14: Please respond = latest May 23 Thanks, Alvin. I fixed the "Protomics" and the "e.g" but don't find = any others. Did you find any others? Posted version has the fixes. =20 Thanks, Eric =20 =20 -------------------------------------------------------------------------= ----- From: mge...@li... = [mailto:mge...@li...] On Behalf Of Alvin = Liu Sent: Tuesday, May 17, 2005 3:10 PM To: mge...@li... Subject: Re: [Mged-misfishie] MISFISHIE draft v14: Please respond = latest May 23 =20 Eric, do a spellcheck, there are still some typos. ----- Original Message -----=20 From: Eric Deutsch=20 To: mge...@li...=20 Cc: Eric Deutsch=20 Sent: Tuesday, May 17, 2005 3:01 PM Subject: [Mged-misfishie] MISFISHIE draft v14: Please respond latest = May 23 =20 Hi everyone, Thank you to all who provided feedback to previous drafts. I have = merged everything into another draft, and here is the result: = http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v14.doc Please read it over and send back any comments to the email list: mge...@li... If you do not belong to this list, your may join or view the = archives at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ Please do send comments if there are any more or you haven't chimed = in yet as soon as you can. Please get all comments back to me on Monday = May 23 at the latest. If you have questions, please let me know or post = the questions to the mailing list. I do hope to submit the revised = manuscript to Genome Biology fairly soon, but there is still not = complete consensus on the minimum requirements, so I'm keeping the = comment period open a little longer. Many thanks to all who have contributed and participated. I = apologize to those who receive this message multiple times via the = various distribution lists. I encourage you to forward this message to = additional people whom you think should see the manuscript but may not = be on these distribution lists. Note that you can view the current = members list at the SourceForge URL above. I will also be sending this = manuscript to a few more members of the community not on these lists, = seeking input from them. If you would like to suggest additional people = and would rather have me send them an email, let me know. Regards, Eric |
From: Eric D. <ede...@sy...> - 2005-05-18 05:29:34
|
Thanks, Alvin. I fixed the "Protomics" and the "e.g" but don't find any others. Did you find any others? Posted version has the fixes. =20 Thanks, Eric =20 =20 _____ =20 From: mge...@li... [mailto:mge...@li...] On Behalf Of Alvin Liu Sent: Tuesday, May 17, 2005 3:10 PM To: mge...@li... Subject: Re: [Mged-misfishie] MISFISHIE draft v14: Please respond latest May 23 =20 Eric, do a spellcheck, there are still some typos. ----- Original Message -----=20 From: Eric Deutsch <mailto:ede...@sy...> =20 To: mge...@li...=20 Cc: Eric Deutsch <mailto:ede...@sy...> =20 Sent: Tuesday, May 17, 2005 3:01 PM Subject: [Mged-misfishie] MISFISHIE draft v14: Please respond latest May 23 =20 Hi everyone, Thank you to all who provided feedback to previous drafts. I have merged everything into another draft, and here is the result: =09 http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v14.doc <http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v14.doc>=20 Please read it over and send back any comments to the email list: mge...@li... <mailto:mge...@li...>=20 If you do not belong to this list, your may join or view the archives at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ <http://lists.sourceforge.net/lists/listinfo/mged-misfishie/>=20 Please do send comments if there are any more or you haven't chimed in yet as soon as you can. Please get all comments back to me on Monday May 23 at the latest. If you have questions, please let me know or post the questions to the mailing list. I do hope to submit the revised manuscript to Genome Biology fairly soon, but there is still not complete consensus on the minimum requirements, so I'm keeping the comment period open a little longer. Many thanks to all who have contributed and participated. I apologize to those who receive this message multiple times via the various distribution lists. I encourage you to forward this message to additional people whom you think should see the manuscript but may not be on these distribution lists. Note that you can view the current members list at the SourceForge URL above. I will also be sending this manuscript to a few more members of the community not on these lists, seeking input from them. If you would like to suggest additional people and would rather have me send them an email, let me know. Regards, Eric |
From: Alvin L. <aliu@u.washington.edu> - 2005-05-17 22:09:54
|
MISFISHIE draft v14: Please respond latest May 23Eric, do a spellcheck, = there are still some typos. ----- Original Message -----=20 From: Eric Deutsch=20 To: mge...@li...=20 Cc: Eric Deutsch=20 Sent: Tuesday, May 17, 2005 3:01 PM Subject: [Mged-misfishie] MISFISHIE draft v14: Please respond latest = May 23 Hi everyone, Thank you to all who provided feedback to previous drafts. I have = merged everything into another draft, and here is the result: http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v14.doc Please read it over and send back any comments to the email list: mge...@li... If you do not belong to this list, your may join or view the archives = at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ Please do send comments if there are any more or you haven't chimed in = yet as soon as you can. Please get all comments back to me on Monday = May 23 at the latest. If you have questions, please let me know or post = the questions to the mailing list. I do hope to submit the revised = manuscript to Genome Biology fairly soon, but there is still not = complete consensus on the minimum requirements, so I'm keeping the = comment period open a little longer. Many thanks to all who have contributed and participated. I apologize = to those who receive this message multiple times via the various = distribution lists. I encourage you to forward this message to = additional people whom you think should see the manuscript but may not = be on these distribution lists. Note that you can view the current = members list at the SourceForge URL above. I will also be sending this = manuscript to a few more members of the community not on these lists, = seeking input from them. If you would like to suggest additional people = and would rather have me send them an email, let me know. Regards, Eric |
From: Eric D. <ede...@sy...> - 2005-05-17 22:01:10
|
Hi everyone, Thank you to all who provided feedback to previous drafts. I have merged everything into another draft, and here is the result: http://scgap.systemsbiology.net/standards/misfishie/MISFISHIE_v14.doc Please read it over and send back any comments to the email list: mge...@li... If you do not belong to this list, your may join or view the archives at http://lists.sourceforge.net/lists/listinfo/mged-misfishie/ Please do send comments if there are any more or you haven't chimed in yet as soon as you can. Please get all comments back to me on Monday May 23 at the latest. If you have questions, please let me know or post the questions to the mailing list. I do hope to submit the revised manuscript to Genome Biology fairly soon, but there is still not complete consensus on the minimum requirements, so I'm keeping the comment period open a little longer. Many thanks to all who have contributed and participated. I apologize to those who receive this message multiple times via the various distribution lists. I encourage you to forward this message to additional people whom you think should see the manuscript but may not be on these distribution lists. Note that you can view the current members list at the SourceForge URL above. I will also be sending this manuscript to a few more members of the community not on these lists, seeking input from them. If you would like to suggest additional people and would rather have me send them an email, let me know. Regards, Eric |