From: Mark G. <mg...@us...> - 2003-01-28 23:11:52
|
Update of /cvsroot/gmod/apollo/doc In directory sc8-pr-cvs1:/tmp/cvs-serv21400 Modified Files: fixed-bugs Log Message: Index: fixed-bugs =================================================================== RCS file: /cvsroot/gmod/apollo/doc/fixed-bugs,v retrieving revision 1.85 retrieving revision 1.86 diff -C2 -d -r1.85 -r1.86 *** fixed-bugs 28 Jan 2003 22:29:02 -0000 1.85 --- fixed-bugs 28 Jan 2003 23:11:48 -0000 1.86 *************** *** 2,210 **** Bugs fixed after 1/28/02 (bugs fixed before then have already been removed) --------------------------------------------------------------------------- - EDE:48) a)On exon delete in EDE the stop codon is not being recalculated - (FS.setTranslationendFromStart) - b) nor is the frame shift being recalculated in the glyph - This can be seen by comparing it with exon deletion in main window - (exon #2 in CG1113 for instance) - this is related to #40 - EDE:40) Diagnostic message on deleting exon with stop codon in ede on - CG1158 example.xml: - CG1158-RA has no feature containing 1126705 (1125864-1126620) - CG1158-RA has no feature containing 1126705 (1125864-1126620) - EDE:41) When doing an external exon selection the figure updates but - selection goes away in the ede sequence [after forcing a repaint - see #39] where it works fine with external transcript selection. - EDE:39) Repaint not happening on external selection: if you select a - transcript in the main window, the glyph changes but the ede selected - sequence does not change its selection. If you force a repaint the - selection suddenly changes. - EDE:42) Create exon creates the exon in the first transcript only, no - matter which transcript was actually clicked on. - This now works fine (or at least i cant recreate it), so perhaps one - of the other bug fixes fixed this as well (?) - EDE:46) Change wording of move buttons - EDE:5) Sometimes ede/highlight bar does not follow selection when a new - gene is selected in the main window. This seems to be ok initially and - goes wrong when bringing up the EDE for the second time. - Bev: I don't want it to move. - MG: per bevs request should we not have ede listen to external - selection or even better just only listen to selections in the - area of the gene it cares about so it can still change on exons within - gene. - fixed by new selection paradigm - need feedback from harvard - EDE:44) Change "Export sequence to clipboard" to "Sequence" with seq window like main window - EDE:24) Problem with frames and stop codons reported by Chris S: - Scaffold - ~bio-adm/gadflyUexport/gbunits/xml/211000022278190.xml - Gene CG40269 - I deleted the 3' (3rd) exon of this model and wanted to reset the stop on - the 2nd exon by extending its end with the exon detail editor. - The expected stop for this frame is at 278670. However Apollo inserted a - stop codon at 278,496. The gene model shows a long ORF that does not - actually exist in any single reading frame. It looks as if Apollo somehow - thinks the exon is in a different frame than it is.[high] - Mark:So I downloaded the file but unfortunately it is no longer in the 3 - exon state. CG40269 is now a 2 exon model and interestingly enough it - stops at 278671 not 278670. I was wondering if you knew of another - example or if there was someway of doctoring this entry to get it back - in the state where it was making the error. - I attempted my own doctoring. I split the second exon which made the - stop go on the 3rd exon. I then deleted the 3rd exon which made the - transcript have a start with no stop. I then started dragging the end - of the 2nd exon. What I found is no matter where I dragged it would - put a stop in of just one base even though there was no stop codon in - the translation. It would put in the proper stop once I dragged it far - enough to get a proper stop, but until that point it just puts a stop - at the end no matter what. This seems improper to me. Or is the idea - that if there is no stop codon then make the last base a stop codon? - Suzi do you know about this? - Took out micheles adding of fake stop at end. Its unclear if this is identical - to the problem chris reported but he looked at it and thought it was. - EDE:26)Bev:Another big problem to us is that you can only bring up one EDE. We - don't like having the EDE view move as you select different exons. I - much prefer being able to open more than one EDE and have each stay at - the spot I am focusing on.[high] - Suzi: one factor in changing to just one was memory issues as several - sequence objects are added. - MG: looked into the additional mem usage. After loading is done and gc - it only adds 2M. - MG: One issue with this is selection. Multiple windows should probably not - respond to external selection. Nomis idea was to have only the most recent ede respond - to selection, just like sequence window. John and Chris agreed that would be best. - EDE:12) When you make the exon editor window bigger, the sequence panel doesn't - always expand. - Bev: This still seems spotty. - Nomi: This is most problematic on Solaris. - EDE:42) Create exon creates the exon in the first transcript only, no - matter which transcript was actually clicked on. - xEDE:22) middle mouse button doesn't seem to be working - Should be able to use third (middle) mouse button to select "Export - sequence to clipboard" > select copy genomic sequence > test by - BLASTNing against predicted genes Only enabled if glyph is enabled - - glyph disabled if click on protein sequence above nucleotide sequence - This is a menu item on right click (not middle) and it appears to be - working ok. - EDE:21) no mutual selection between exon view and main view - (true model view controller would solve this) - There is now mutual selection. If we go to one ede per gene we - probably want to disable some of it. - EDE:15)x- Shouldn't hash-highlight from previous find sequence go away when you - do another find? - appears to work now. - EDE:27)Bev:I really liked the fact that in old Apollo you could select that last - exon (or any spot on a transcript) in the main viewer, open BOFE, and - have it centered around that location. I do not like the fact that any - time you open it now it is always fixed at the 5' end of the - transcript. It's very inconvenient.[medium] - This is actually the same as 7 - woops. Actually slightly different - than 7. 7 is for selection (which bev might oppose), this is when you - initially bring up. - EDE:7) Would be nice if EDE scrolled to exon selected in main window. - EDE:34) Delete exon in ede right click menu doesnt work right. The exon - goes away in the main window but remains in the ede. - EDE29)Bev:I don't like the fact that when the EDE is open, you can't do - operations such as gene merge in the main viewer. With the EDE open, - you can't select a second gene. I think this has something to do with - the EDE and the main viewer being more coupled in the new version to - allow the EDE view to shift as you select different exons in the main - viewer, etc. I personally don't like that idea and would like to see it - work as in old apollo. - This seems like the same issue as 2[high] - EDE16)x- Clicking exons in figure doesn't always scroll to them in detail view - [can't repeat this error] - i think this got fixed. - EDE13) Merge 5' exon not working? (reported by Suzi) - EDE:4) When selecting different transcripts in the main window the EDE does - not follow the selection (where vice versa does work) - Bev: I don't like this feature anyway - EDE:2) The EDE sometimes makes it so individual exons in the main window - can't get selected. Apon selecting one the whole transcript - selects. This is sporatic. - Bev:To make changes in the annotation zone you would have to close the EDE - first which is very inconvenient and disruptive.[medium] - Bev: This is definately annoying but sporadic. - xEDE:25) (from Bev) If you add - transcripts or change transcripts in the main viewer, the EDE does not - get updated. This is a show stopper. It is very important that the EDE - get a new line of sequence and a new structure every time you add a - transcript and that the structures agree with what is in the main - viewer.[high] - I looked into this further. So initially ede reflects adding and changing trans in - main window. Its only after something else has happened that it goes - awry. So after a failed "delete selection"(in main window), then things fail to get - reflected, so the delete selection bug(see Annotation editing problems - above) should be worked on first. - Fixed delete bug and it exposed a few more problems with redraw and - nulled out genes that were then fixed as well. - xEDE:31) (or 11e) This relates to 11, 8, and 18. (could be 11e i guess). In - example.xml on CG1161 after doing a few 3' exon merges and getting - the messages of 11b and 11d another exon merge can bring the message: - FeatureSet.setPhases: CG1161-RA can't set phases--some features undefined tss=1180904 tes=1182464 - and the start and stop are lost. This is the only way I have found to - replicate bugs 8 and 18 (which I think are the same bug). - EDE:18) figure not updating stop codon when changed[med] -> see 31 - EDE:11) If I merge all the exons of CG1161-RA(example.xml) using merge 5' errors(from - right to left): - 11a) after 2nd merge if i then click on the sequence of where the 3rd - exon was I get this message: (same after 3rd merge) - ERROR: Didn't find CG1161-RA exon 3 in set CG1161-RA - 11b) also after 2nd merge glyph and main window shows exons merged, EDE - sequence still shows the intron, same after 3rd merge (but 1st merge ok) - 11c) apon doing the last merge I - get a popup message saying it wants to remove surrounding exon 4 and - then I get an exception: - java.lang.NullPointerException - at apollo.gui.AnnotationEditor.mergeFeatures(AnnotationEditor.java:2060) - at apollo.gui.AnnotationEditor.mergeFeatures(AnnotationEditor.java:1980) - at apollo.gui.detailviewers.exonviewer.BaseEditorPanel$RightClickActionListener.actionPerformed(BaseEditorPanel.java:874) - at javax.swing.AbstractButton.fireActionPerformed(AbstractButton.java:1450) - 11d) in doing the 2nd merge on CG1113-RA (2nd and 3rd exon already - succefully merged trying to merge with 1st exon): - Could not delete CG1113-RA exon 2 from CG1113-RA because drawable was - not found - and it does not merge the exons. - 11e) see 31 - 11f) after 1st merge the newly merged end exon now enables the merge - with 3' exon which shouldnt be enabled, and the original end exon did - not have it enabled - EDE:10) Transcript figure doesnt redraw itself properly on exon merging[high] - EDE:8) Stop codon not showing up in glyph after exon merge.[med] -> see 31 - - Save broken. Gives following exception: - caught exception committing XML - java.util.Vector - java.lang.ClassCastException: java.util.Vector - at apollo.dataadapter.GAMEAdapter.writeProperties(GAMEAdapter.java:1104) - at apollo.dataadapter.GAMEAdapter.writeAnnotation(GAMEAdapter.java:370) - at apollo.dataadapter.GAMEAdapter.writeAnnotations(GAMEAdapter.java:345) - at apollo.dataadapter.GAMEAdapter.writeXML(GAMEAdapter.java:300) - at apollo.dataadapter.GAMEAdapter.commitChanges(GAMEAdapter.java:265) - at apollo.dataadapter.GAMEAdapterGUI.doOperation(GAMEAdapterGUI.java:104) - at org.bdgp.swing.widget.DataAdapterChooser.doCommitWithExceptions(DataAdapterChooser.java:355) - at org.bdgp.swing.widget.DataAdapterChooser.doCommit(DataAdapterChooser.java:408) - at apollo.gui.menus.FileMenu$SaveAdapterChooser.doCommit(FileMenu.java:282) - at org.bdgp.swing.widget.DataAdapterChooser$CommitRunnable.run(DataAdapterChooser.java:132) - at java.lang.Thread.run(Thread.java:484) - - EDE:23) User manual mentions "Create tiny exon" menu option--not yet - implemented. Need? Bev:Yes [high] - Instructions from user manual: - 1. Right click on a base in an intron. (You cannot create an exon - outside of the transcript in the exon editor.)<BR> - 2. Left click on 'Create tiny exon'.<BR> - 3. A three base exon is made within the intron. The base that you - clicked on is the middle base in the three base exon. Currently just - makes a one base exon. Is there a reason it has to be 3? Is the - minimum size of an exon 3? - - EDE:3) Make intron seems all wrong. EDE makes an intron all the way to the - end of the transcript. The main window just puts in a one base intron - which im guessing is what the ede should be doing. - Bev:This was also a problem in the old apollo that was very limiting.[high] - - xEDE:6) Sometimes glyph does not update with selecting different transcript - (sequence part) in the EDE.[high] - - xEDE:1) When the ends of an exon are adjusted the translated sequence in the - exon in the transcript in the main window does not recompute but stays - with the old translation. (The sequence window does show the new translation)[high] - x If you try to load a new dataset and the load fails, you lose the display of the old dataset (it's already been cleared out). --- 2,5 ---- *************** *** 2324,2325 **** --- 2119,2326 ---- data set remain on new load. Either the selection needs to be cleared or the whole window should get disposed on new load. + + EDE:48) a)On exon delete in EDE the stop codon is not being recalculated + (FS.setTranslationendFromStart) + b) nor is the frame shift being recalculated in the glyph + This can be seen by comparing it with exon deletion in main window + (exon #2 in CG1113 for instance) + this is related to #40 + EDE:40) Diagnostic message on deleting exon with stop codon in ede on + CG1158 example.xml: + CG1158-RA has no feature containing 1126705 (1125864-1126620) + CG1158-RA has no feature containing 1126705 (1125864-1126620) + EDE:41) When doing an external exon selection the figure updates but + selection goes away in the ede sequence [after forcing a repaint + see #39] where it works fine with external transcript selection. + EDE:39) Repaint not happening on external selection: if you select a + transcript in the main window, the glyph changes but the ede selected + sequence does not change its selection. If you force a repaint the + selection suddenly changes. + EDE:42) Create exon creates the exon in the first transcript only, no + matter which transcript was actually clicked on. + This now works fine (or at least i cant recreate it), so perhaps one + of the other bug fixes fixed this as well (?) + EDE:46) Change wording of move buttons + EDE:5) Sometimes ede/highlight bar does not follow selection when a new + gene is selected in the main window. This seems to be ok initially and + goes wrong when bringing up the EDE for the second time. + Bev: I don't want it to move. + MG: per bevs request should we not have ede listen to external + selection or even better just only listen to selections in the + area of the gene it cares about so it can still change on exons within + gene. + fixed by new selection paradigm - need feedback from harvard + EDE:44) Change "Export sequence to clipboard" to "Sequence" with seq window like main window + EDE:24) Problem with frames and stop codons reported by Chris S: + Scaffold - ~bio-adm/gadflyUexport/gbunits/xml/211000022278190.xml + Gene CG40269 + I deleted the 3' (3rd) exon of this model and wanted to reset the stop on + the 2nd exon by extending its end with the exon detail editor. + The expected stop for this frame is at 278670. However Apollo inserted a + stop codon at 278,496. The gene model shows a long ORF that does not + actually exist in any single reading frame. It looks as if Apollo somehow + thinks the exon is in a different frame than it is.[high] + Mark:So I downloaded the file but unfortunately it is no longer in the 3 + exon state. CG40269 is now a 2 exon model and interestingly enough it + stops at 278671 not 278670. I was wondering if you knew of another + example or if there was someway of doctoring this entry to get it back + in the state where it was making the error. + I attempted my own doctoring. I split the second exon which made the + stop go on the 3rd exon. I then deleted the 3rd exon which made the + transcript have a start with no stop. I then started dragging the end + of the 2nd exon. What I found is no matter where I dragged it would + put a stop in of just one base even though there was no stop codon in + the translation. It would put in the proper stop once I dragged it far + enough to get a proper stop, but until that point it just puts a stop + at the end no matter what. This seems improper to me. Or is the idea + that if there is no stop codon then make the last base a stop codon? + Suzi do you know about this? + Took out micheles adding of fake stop at end. Its unclear if this is identical + to the problem chris reported but he looked at it and thought it was. + EDE:26)Bev:Another big problem to us is that you can only bring up one EDE. We + don't like having the EDE view move as you select different exons. I + much prefer being able to open more than one EDE and have each stay at + the spot I am focusing on.[high] + Suzi: one factor in changing to just one was memory issues as several + sequence objects are added. + MG: looked into the additional mem usage. After loading is done and gc + it only adds 2M. + MG: One issue with this is selection. Multiple windows should probably not + respond to external selection. Nomis idea was to have only the most recent ede respond + to selection, just like sequence window. John and Chris agreed that would be best. + EDE:12) When you make the exon editor window bigger, the sequence panel doesn't + always expand. + Bev: This still seems spotty. + Nomi: This is most problematic on Solaris. + EDE:42) Create exon creates the exon in the first transcript only, no + matter which transcript was actually clicked on. + xEDE:22) middle mouse button doesn't seem to be working + Should be able to use third (middle) mouse button to select "Export + sequence to clipboard" > select copy genomic sequence > test by + BLASTNing against predicted genes Only enabled if glyph is enabled + - glyph disabled if click on protein sequence above nucleotide sequence + This is a menu item on right click (not middle) and it appears to be + working ok. + EDE:21) no mutual selection between exon view and main view + (true model view controller would solve this) + There is now mutual selection. If we go to one ede per gene we + probably want to disable some of it. + EDE:15)x- Shouldn't hash-highlight from previous find sequence go away when you + do another find? - appears to work now. + EDE:27)Bev:I really liked the fact that in old Apollo you could select that last + exon (or any spot on a transcript) in the main viewer, open BOFE, and + have it centered around that location. I do not like the fact that any + time you open it now it is always fixed at the 5' end of the + transcript. It's very inconvenient.[medium] + This is actually the same as 7 - woops. Actually slightly different + than 7. 7 is for selection (which bev might oppose), this is when you + initially bring up. + EDE:7) Would be nice if EDE scrolled to exon selected in main window. + EDE:34) Delete exon in ede right click menu doesnt work right. The exon + goes away in the main window but remains in the ede. + EDE29)Bev:I don't like the fact that when the EDE is open, you can't do + operations such as gene merge in the main viewer. With the EDE open, + you can't select a second gene. I think this has something to do with + the EDE and the main viewer being more coupled in the new version to + allow the EDE view to shift as you select different exons in the main + viewer, etc. I personally don't like that idea and would like to see it + work as in old apollo. + This seems like the same issue as 2[high] + EDE16)x- Clicking exons in figure doesn't always scroll to them in detail view + [can't repeat this error] - i think this got fixed. + EDE13) Merge 5' exon not working? (reported by Suzi) + EDE:4) When selecting different transcripts in the main window the EDE does + not follow the selection (where vice versa does work) + Bev: I don't like this feature anyway + EDE:2) The EDE sometimes makes it so individual exons in the main window + can't get selected. Apon selecting one the whole transcript + selects. This is sporatic. + Bev:To make changes in the annotation zone you would have to close the EDE + first which is very inconvenient and disruptive.[medium] + Bev: This is definately annoying but sporadic. + xEDE:25) (from Bev) If you add + transcripts or change transcripts in the main viewer, the EDE does not + get updated. This is a show stopper. It is very important that the EDE + get a new line of sequence and a new structure every time you add a + transcript and that the structures agree with what is in the main + viewer.[high] + I looked into this further. So initially ede reflects adding and changing trans in + main window. Its only after something else has happened that it goes + awry. So after a failed "delete selection"(in main window), then things fail to get + reflected, so the delete selection bug(see Annotation editing problems + above) should be worked on first. + Fixed delete bug and it exposed a few more problems with redraw and + nulled out genes that were then fixed as well. + xEDE:31) (or 11e) This relates to 11, 8, and 18. (could be 11e i guess). In + example.xml on CG1161 after doing a few 3' exon merges and getting + the messages of 11b and 11d another exon merge can bring the message: + FeatureSet.setPhases: CG1161-RA can't set phases--some features undefined tss=1180904 tes=1182464 + and the start and stop are lost. This is the only way I have found to + replicate bugs 8 and 18 (which I think are the same bug). + EDE:18) figure not updating stop codon when changed[med] -> see 31 + EDE:11) If I merge all the exons of CG1161-RA(example.xml) using merge 5' errors(from + right to left): + 11a) after 2nd merge if i then click on the sequence of where the 3rd + exon was I get this message: (same after 3rd merge) + ERROR: Didn't find CG1161-RA exon 3 in set CG1161-RA + 11b) also after 2nd merge glyph and main window shows exons merged, EDE + sequence still shows the intron, same after 3rd merge (but 1st merge ok) + 11c) apon doing the last merge I + get a popup message saying it wants to remove surrounding exon 4 and + then I get an exception: + java.lang.NullPointerException + at apollo.gui.AnnotationEditor.mergeFeatures(AnnotationEditor.java:2060) + at apollo.gui.AnnotationEditor.mergeFeatures(AnnotationEditor.java:1980) + at apollo.gui.detailviewers.exonviewer.BaseEditorPanel$RightClickActionListener.actionPerformed(BaseEditorPanel.java:874) + at javax.swing.AbstractButton.fireActionPerformed(AbstractButton.java:1450) + 11d) in doing the 2nd merge on CG1113-RA (2nd and 3rd exon already + succefully merged trying to merge with 1st exon): + Could not delete CG1113-RA exon 2 from CG1113-RA because drawable was + not found + and it does not merge the exons. + 11e) see 31 + 11f) after 1st merge the newly merged end exon now enables the merge + with 3' exon which shouldnt be enabled, and the original end exon did + not have it enabled + EDE:10) Transcript figure doesnt redraw itself properly on exon merging[high] + EDE:8) Stop codon not showing up in glyph after exon merge.[med] -> see 31 + + Save broken. Gives following exception: + caught exception committing XML + java.util.Vector + java.lang.ClassCastException: java.util.Vector + at apollo.dataadapter.GAMEAdapter.writeProperties(GAMEAdapter.java:1104) + at apollo.dataadapter.GAMEAdapter.writeAnnotation(GAMEAdapter.java:370) + at apollo.dataadapter.GAMEAdapter.writeAnnotations(GAMEAdapter.java:345) + at apollo.dataadapter.GAMEAdapter.writeXML(GAMEAdapter.java:300) + at apollo.dataadapter.GAMEAdapter.commitChanges(GAMEAdapter.java:265) + at apollo.dataadapter.GAMEAdapterGUI.doOperation(GAMEAdapterGUI.java:104) + at org.bdgp.swing.widget.DataAdapterChooser.doCommitWithExceptions(DataAdapterChooser.java:355) + at org.bdgp.swing.widget.DataAdapterChooser.doCommit(DataAdapterChooser.java:408) + at apollo.gui.menus.FileMenu$SaveAdapterChooser.doCommit(FileMenu.java:282) + at org.bdgp.swing.widget.DataAdapterChooser$CommitRunnable.run(DataAdapterChooser.java:132) + at java.lang.Thread.run(Thread.java:484) + + EDE:23) User manual mentions "Create tiny exon" menu option--not yet + implemented. Need? Bev:Yes [high] + Instructions from user manual: + 1. Right click on a base in an intron. (You cannot create an exon + outside of the transcript in the exon editor.)<BR> + 2. Left click on 'Create tiny exon'.<BR> + 3. A three base exon is made within the intron. The base that you + clicked on is the middle base in the three base exon. Currently just + makes a one base exon. Is there a reason it has to be 3? Is the + minimum size of an exon 3? + + EDE:3) Make intron seems all wrong. EDE makes an intron all the way to the + end of the transcript. The main window just puts in a one base intron + which im guessing is what the ede should be doing. + Bev:This was also a problem in the old apollo that was very limiting.[high] + + xEDE:6) Sometimes glyph does not update with selecting different transcript + (sequence part) in the EDE.[high] + + xEDE:1) When the ends of an exon are adjusted the translated sequence in the + exon in the transcript in the main window does not recompute but stays + with the old translation. (The sequence window does show the new translation)[high] + |