[Animl-develop] Plate reader data in AnIML

[Amanda C. <af...@ab...>]

Dear AnIMLers,

I'm interested in the application of XML standards to plate reader data, 
because we have some data that we would like to make available in XML, 
and I was curious to see if AnIML would be the way to go for this. I 
notice that in the minutes of your meeting on 1st April 2005 there was 
the comment "Burkhard noted that if it is not already there, one 
requirement should be that AnIML does not become obsolete when new 
analytical techniques are developed. For example, 96 well plate readers".

So I thought I'd put together an example of how it might look so I could 
get a feel for using this as opposed to creating my own XML format for 
my own purposes. I've created a few files:

http://users.aber.ac.uk/afc/work/platereaders/20041208-2572.victor.animl
http://users.aber.ac.uk/afc/work/platereaders/PlateReader_base_instance_document.atid

Of course, they are not by any means detailed or complete, just enough 
to get a feel for the pros and cons of using the AnIML format to 
represent my data. I now have a few questions about AnIML, would it be 
appropriate to ask these here?

1) Do you have plans to include microtitre plate reader data in future 
(and if so, would it look something like this, am I on the right lines)? 
I think the main issue I had was trying to record pairs of values in a 
vector. For each well within a 96 well plate we have many readings, and 
for each reading there is a timestamp I'd like to record. For this I 
used a Page for each well, and a VectorSet of 2 Vectors within each Page 
(one Vector for the readings and one Vector for the corresponding 
timestamps).

2) I think I don't yet fully understand the main purpose of AnIML. That 
is, as I see it there's a difference between automatically collected 
data that is created by the machine (plate reader, or whatever, 
including the parameters that were used to set up that run of the 
machine), and human input data that is describing what the data is to be 
used for and how. A difference between data that relates to the 
measurements and data that relates to the purpose of the expt. I feel 
the former should be required, and the latter should be optional. At the 
moment AnIML seems to require both.
    For example: if I describe the samples used in an experiment I'm 
required to describe their role. And if I describe a Vector, I must say 
whether it contains a dependent or independent variable, and what its 
plot use is to be. The plate reader machine can hardly know this, though 
a human would. So I am suprised that these attributes are "required" 
rather than "optional".

3) Are there tools to allow validation against the atid document? At the 
moment this seems to be like a meta-schema, and I can't use it for 
validating my documents.


Then I have various questions about technical parts of AnIML that I 
couldn't find answers to on the website. I apologise if they have 
already been answered elsewhere and I haven't come across it yet.

1) float vs float32/64
In the *technique.atid documents "float" is used: VectorBlueprint 
name="Readings" type="float"
In the animl-technique.xsd docs float is used for the 
VectorTypeNames/AllTypeNames
In the animl-core.xsd VectorValueType can be float32 or float64 (mapping 
to xs:float and xs:double respectively)

2) Namespaces: will they be used in future? It makes AnIML much less 
useful without them.

3) Filenames with spaces (such as animl-core 1.04.xsd) sometimes cause 
problems (for eaxmple, if you have namespaces and want to specify 
multiple schemaLocations for the various namespaces), and I think it 
would be better to avoid spaces in filenames if possible.

4) I am partly confused about the distinction between modality and using 
maxOccurs/minOccurs in the atid docs. Why does animl-technique allow 
both maxOccurs and modality? what happens if they conflict? And why is 
minOccurs not allowed?

5) What value should I use for values that weren't recorded? (NaN? N/A?) 
I have many of these.

6) Which are the versions of animl-core.xsd and animl-technique.xsd that 
I should be using, and do they have any relationship to the compulsory 
'version="1.0"' attribute of animl-core or 1.1 of animl-technique?

Thank you in advance for any help and advice.


Amanda

-- 
Amanda Clare  http://users.aber.ac.uk/afc/
Tel: +44 (0)1970 622410  Fax: +44 (0)1970 628536
Dept. of Computer Science, University of Wales, Aberystwyth, SY23 3DB