Skip to content
Snippets Groups Projects
Commit cf47b0aa authored by edgrif's avatar edgrif
Browse files

moved to 2006

parent 7c9c2c5a
No related branches found
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
Showing
with 0 additions and 1226 deletions
ZMAP_LACE_PROJECT/zmap_lace.2006_06_21 deleted 100755 → 0
+ 0
− 150
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Wens 21st June 2006
Attendees: th, jla1, kj2, jgrg, edgrif
1) otterlace + zmap
===================
Phase 1
-------
Strategy is to have Liz test for now as she is an experienced user
and we will be able to respond to her suggestions quickly.
For this phase lace is set up so that xace/fmap and zmap can be
used side by side for comparison.
ZMap issues arising:
locus names missing
bumping (maybe solved, not tested)
3 frame translation (roy)
spandit (fixed, not tested)
keyboard shortcuts: check against spandit
otterlace issues:
needs new tk for clone_descriptions
'annotated' should be a status, not text
Phase 2
-------
Extend usage to other experienced but more varied users, Charlie and
Gavin suggested as good candidates.
Genefinding
-----------
Some annotators make considerable usage of the built in genefinder in
fmap, acedb can export these features, zmap needs some additional drawing
routines to display them.
Linux/Mac builds only
---------------------
Currently because of /usr/local issues zmap cannot be built on the Alpha,
unless there is some big change in requirements zmap will only be supported
on Linux and the Mac.
Suggestion/Bug tracking
-----------------------
Both acedb and zmap use the Request Tracker system, send emails to "acedb-bug"
and "zmap" respectively.
Queues should also be set up for vega (with external connection) and anacode.
Xace usage
----------
Currently there is still a small and decreasing need for write access via
xace to allow comments/notes to be added to some acedb objects. This will
need to disappear when xace is removed. James is working on this.
ZMap development issues
-----------------------
- multiple segments:
A lot of discussion about how this would work, there are several different
requirements and several ways this could work.
James would like to keep the way lace extracts data (i.e. as a single
contiguous segment) the same as he has spent much time producing code
that does just that in a reliable way. Ed said that zmap could extract
regions from this in a way that would be useful to the annotator.
When it comes to multiple alignments it may be best for zmap to retrieve
them from several separate otterlace databases.
This needs much more discussion/refinement.
- navigation in zmap:
Issue is gestures to allow easy navigation, James suggested a 'back' button
(like browser) going back to previous zoom and/or position ?
Jen would like to be able to skip from exon to exon, Ed suggested overview
of gene with second pane that shows just one exon but can skip to the next.
Otterlace
---------
- External rollout:
last non http removed
ssh access first
SSL access externally later
Jen needs by August
- DAS issues:
otterlace can display das sources on clone
most of das is on chromosome coordinates
Jen wants to see ditags
James/Tim will work out a solution for this by time of next meeting
- Ensembl
Conservation plots, when are they coming?
TH to check: Jen says Adam S's plan is excellent.
New schema:
end of July
more intelligent comparison of output
Realign_offtrack_genes
producing random shifts
use previous versions of annotation to find large shifts
either
by genome genome alignment
by tracking missing evidence
new schema will address this
mouse is priority, but also problem in human (CCDS, merge)
==============================================================================
ZMAP_LACE_PROJECT/zmap_lace.2006_07_12 deleted 100755 → 0
+ 0
− 181
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Wens 12th July 2006
Attendees: th, jla1, kj2, jgrg, edgrif
1) otterlace + zmap progress
============================
new features
------------
- display of locus names, supposed to be done.
- bumping, supposed to be done
both these items require new version, this highlighted
the following points:
1) zmap team must deliver regular updates to anacode in a controlled
and visible way (via RT ticket pointing James to latest release).
2) anacode team must then release new zmap + their code quickly.
3) havana guinea pig testers need to ensure zmap + anacode teams
appreciate urgency/priority of requests, e.g. "I can't work without
this item".
- liz's treeview editing requests
Liz needs to see extra data about objects when editting,
James/Ed agreed that the correct way to do this is to have zmap
communicate the selected object id to lace and then for lace to
display the information in a separate window. This would mirror
exactly the existing situation whereby Liz selects and object
in fmap and its details are displayed in the acedb treeview display.
This is a priority, since testing of zmap is blocked until this is
done. Roy has done his bit so that feature information is passed
back to lace, James to implement lace code in the next week.
- RT queues
We need zmap/anacode queues linked so tickets can be transferred
between the two.
James has started process of setting up RT - will be default for
anacode email. anacode-people for people.
RT tickets: can't see other people's by default? Would be better
if you can.
- match/alignment display
kerstin would like to be able to cluster/join up alignments that clearly
are all part of the same match. Ed says this will require a more generalised
mechanism than exists in acedb currently. We could use the Join_homols tag
with an Int field:
Join_homols nnnn // where nnnn specfies an upper threshold in align
// gap for joining two consecutive homols of the
// same match.
- blixem
search for sequence string? Horrible code, so just simple searching [ed]
fetching sequences for blixem: pfetch quite fast now, but would be
faster if pfetch cache? either in acedb or separate []
- acedb compilation [ed]:
compiling to gtk2 - now working
building for mac/linux; universal binaries soon.
- external users:
requires version checking [james]
under apache can run multiple servers, so can handle version transition.
- selenocysteine/alternative genetic codes :
ed will check acedb code and make sure zmap can do the alternative
translations.
james will implement selenocysteine in new schema
- zmap/otterlace demo
In one weeks time: Wednesday 19th.
- Zmap/otterlace new guinea pigs
Charlie has been using it but stopped because he needs a bit of
help in orientation...Ed/Roy need to help here.
Gavin would also be good but needs the Gene Finder Features,
Ed to expedite this...but will check with Kerstin/Gavin to
make sure zmap provides the right display.
- multiple alignments
Ed said he has done more work on the flexibility of zmap to
display these but is still unsure (as are we all ??) about
exactly what sort of display is required. Zmap plan is to
produce a straight forward display and get feedback.
2) other matters
================
- DAS source and builds
TH: DAS source for ditag data within 1 week (19th)
TH: will build 19_36 mouse db + 19_36 human db
- genelists were discussed:
looked at genetracker developed by Roger/webteam:
http://intweb.sanger.ac.uk/cgi-bin/utils/genetracker
not being able to add extra columns a major issue.
- extra tags
Discussion about tags on clones and genes indicating status. Need
some long term solution for this. Unclear if this should be in the
gene list tracking system or in otterlace.
For now, using Annotated_remark- annotated and Annotated_remark-
inprogress for clones. Should be possible to set this status from
the interface [James]
Similar tags for states for genes should also be added.
- EMBL dumping
Needs to be possible for annotators to submit a clone on demand [James]
EMBL dumping needs to be aware of gene + clone status tags. If
genes are annotated, but clone is not set to 'annotated' should add a
comment to the header to warn that the annotation for the clone is
currently partial.
Need to be able to lock objects (transcripts) if a tag is set -
i.e. CCDS tag [james says easy to fix]
==============================================================================
--
------------------------------------------------------------------------
| Ed Griffiths, Acedb development, Informatics Group, |
| The Morgan Building, Sanger Institute, Wellcome Trust Genome Campus |
| Hinxton, Cambridge CB10 1HH |
| |
| email: edgrif@sanger.ac.uk Tel: +44-1223-496844 Fax: +44-1223-494919 |
------------------------------------------------------------------------
ZMAP_LACE_PROJECT/zmap_lace.2006_07_26 deleted 100755 → 0
+ 0
− 111
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Wens 26th July 2006
Attendees: th, jla1, kj2, jgrg, edgrif
1) otterlace + zmap progress
============================
new features
------------
- locus names, edgrif and eah agreed this was done but may need to add
names as text to display.
- bumping, the compact algorithm was discussed and the concensus was that
it will not do the job. Instead edgrif will write a variant of this that
does not allow matches to be interleaved. Matches from the same est will
have a background colour painted between them to make it obvious which
matches go with which.
- tickets: The only item to be resolved is that users cannot see all tickets
in a queue, this is essential. edgrif will submit a request to get this done
for acedb, anacode and zmap.
- ests:
add 'intron' where a feature perfectly aligns over a splice, zmap needs to look
through matches and join up ones whose coords are completely correct.
- glyph for single base features: we need scale independent glyphs for a number
of features including the gene finder features. edgrif or rds will implment a
new scale independent drawing type as an add on for foocanvas, this is the most
natural/efficient way to implement this.
- alternative translations: edgrif about half way through code to do this.
- Gene finder features: we can export this data but we need a new drawing type
(see glyph item).
- multiple alignments: edgrif is about a third of the way through implementing
a more general way of displaying arbitrary blocks.
- zmap needs to get sequences that are in acedb from there for blixem, means
issuing a call to the server to get them but this is possible. NOTE that some
sequences are _only_ in acedb.
- we need a "back" button
Builds
------
- gtk2: is now done for acedb and zmap, all code now builds on the same level of
gtk2. BUT there are now no more alpha builds of either, only the mac and linux
are supported.
- universal binaries: requires upgrade of software on our laptop which we will now
do.
Blixem/pfetch
--------------
- blixem: dna searching is NOT DONE, edgrif to expedite.
- pfetch: caching? NOT DONE + where should we cache ?
2) Other matters
================
- client/server: version checking NOT DONE
- alignments for genomic alignments: ??
- das - NOT DONE (th)
- genelists:
done for zebrafish
using mig list for mouse
no human
needs multiple list interfaces
- interface for tags: NOT DONE
- submit button for EMBL dumping
- gene_fragments tag - broken
==============================================================================
ZMAP_LACE_PROJECT/zmap_lace.2006_09_13 deleted 100755 → 0
+ 0
− 160
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Wens 13th September 2006
Attendees: th, jla1, jgrg, edgrif
1) otterlace + zmap progress
============================
new features
------------
- reverse complement code, the code is there and working in the test
otterlace system. BUT there is a bug in the interface to blixem which
results in blixem freezing when used on a reverse complemented sequence.
- for a combination of reasons the reverse complement fix did not make
it through to the annotators for 3 weeks. To try and fix this we will
do the following:
1) with each new release of the lace/zmap system we (a lace
person and a zmap person) will sit with one of the annotators
to check that everything is working
2) publish a list of the changes/fixes for that release
- 3 frame translation is high priority, Ed is working on this and will
have something working by the end of this week so that there can be a
release next week.
- Navigator with clickable locus names and a display of the clone assembly
is high priority, Roy is working on this now and it should be ready for
the next release. We need to allow the user to search for locus names within
the navigator window.
- locus names, as well as having locus names in the navigator, we need them
in a column in the main display.
- bumping, the compact algorithm was discussed and the concensus was that
it will not do the job. Instead edgrif will write a variant of this that
does not allow matches to be interleaved. Matches from the same est will
have a background colour painted between them to make it obvious which
matches go with which.
- the release notes for each release and the code version numbers will be
put in a web page that will be viewable from either the lace or zmap
programs as a main menu item. This will help the annotators to be sure of
which version they are using and of the changes/updates made.
- ests:
add 'intron' where a feature perfectly aligns over a splice, zmap needs to look
through matches and join up ones whose coords are completely correct.
This is still be worked on.
- glyph for single base features: we need scale independent glyphs for a number
of features including the gene finder features. edgrif or rds will implment a
new scale independent drawing type as an add on for foocanvas, this is the most
natural/efficient way to implement this.
This has been implemented and we can display gene finder features, we need the
3 frame translation first before this can be properly used.
- alternative translations: edgrif about half way through code to do this.
- multiple alignments: edgrif is about a third of the way through implementing
a more general way of displaying arbitrary blocks.
- zmap needs to get sequences that are in acedb from there for blixem, means
issuing a call to the server to get them but this is possible. NOTE that some
sequences are _only_ in acedb.
- we need a "back" button
Builds
------
- universal binaries: we have an updated mac laptop with the latest XCode set up
which will allow us to produce the binaries.
Blixem/pfetch
--------------
- blixem: dna searching is NOT DONE, edgrif to expedite.
- pfetch: caching? NOT DONE + where should we cache ?
multiple species
----------------
- James and Ed need to discuss how this will be done. Zmap can display
multiple sequences in one zmap window but there will need to changes to
the supporting infrastructure including:
- should the species go in one or multiple acedb databases
- the lace <-> zmap protocol needs to allow specification of
multiple sequence display.
- lace needs to allow the user to pick multiple species
Acedb
-----
- Treeview has a bug whereby if you click on an item, the window is no longer
preserved. This is because of a fix by Ed and now needs to be refixed.
- need to make sure that both acedb and zmap use the Mac "open" command to
show urls. You can just give the url and "open" will invoke the users default
browser.
2) Other matters
================
sorry, this section is rather sparse...give me more information if you need to.
- das - NOT DONE (th)
- genelists:
done for zebrafish
using mig list for mouse
no human
needs multiple list interfaces
- interface for tags: NOT DONE
- submit button for EMBL dumping
- gene_fragments tag - broken
==============================================================================
ZMAP_LACE_PROJECT/zmap_lace.2006_10_11 deleted 100755 → 0
+ 0
− 140
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Wens 11th October 2006
Attendees: th, jla1, jgrg, kj2, edgrif
1) otterlace + zmap progress
============================
new features
------------
- Bug that led to blixem freezing for reverse complemented sequence
is fixed.
- We will do a complete acedb/zmap build this week and after it is
integrated into lace we will do the following:
1) with each new release of the lace/zmap system we (a lace
person and a zmap person) will sit with one of the annotators
to check that everything is working
2) publish a list of the changes/fixes for that release
- 3 frame translation is done.
- Navigator with clickable locus names and a display of the clone assembly
is high priority. We need to allow the user to search for locus names within
the navigator window. The locus names will also be displayed in the main
column. Roy is working on this now and it should be ready for the next release.
- bumping of homols: Matches from the same est will have a background colour
painted between them to make it obvious which matches go with which and perfect
alignments will have "intron" connectors drawn between them. The intron conector
code still needs to be written. This is a high priority item.
- the release notes for each release and the code version numbers will be
put in a web page that will be viewable from either the lace or zmap
programs as a main menu item. This will help the annotators to be sure of
which version they are using and of the changes/updates made. This release
will be the first time we have this.
- glyph for single base features is done.
- alternative translations: edgrif about half way through code to do this.
- multiple alignments: edgrif is about a third of the way through implementing
a more general way of displaying arbitrary blocks. This is a high priority item.
- zmap needs to get sequences that are in acedb from there for blixem, means
issuing a call to the server to get them but this is possible. NOTE that some
sequences are _only_ in acedb.
- we need a "back" button
- DNA searching in zmap is done.
- There is still a problem with the shutting down of the ace server which
leads to a broken pipe message. edgrif will check that server shuts down
in a timely way.
Builds
------
- universal binaries: we have an updated mac laptop with the latest XCode set up
which will allow us to produce the binaries. edgrif is pursuing both nfs mounts
for automated builds and how to import our standard build system into XCode.
Blixem/pfetch
--------------
- blixem: dna searching is NOT DONE, edgrif to expedite.
- pfetch: caching? NOT DONE + where should we cache ?
multiple species
----------------
- James and Ed need to discuss how this will be done. Zmap can display
multiple sequences in one zmap window but there will need to changes to
the supporting infrastructure including:
- should the species go in one or multiple acedb databases
- the lace <-> zmap protocol needs to allow specification of
multiple sequence display.
- lace needs to allow the user to pick multiple species
Acedb
-----
- Treeview has a bug whereby if you click on an item, the window is no longer
preserved. This is because of a fix by Ed and now needs to be refixed.
- need to make sure that both acedb and zmap use the Mac "open" command to
show urls. DONE
2) Other matters
================
sorry, this section is rather sparse...give me more information if you need to.
- das - NOT DONE (th)
- interface for tags: NOT DONE but Leo working on related code for import of
Ensemble genes that will be directly applicable.
- submit button for EMBL dumping: not done.
==============================================================================
ZMAP_LACE_PROJECT/zmap_lace.2006_11_02 deleted 100755 → 0
+ 0
− 174
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Thurs 2nd November 2006
Attendees: th, jla1, jgrg, kj2, edgrif
1) otterlace + zmap progress
============================
new features
------------
- Bug 14058, blixem reverse coords (reported by Gavin) is fixed.
- Roy sat with Gavin (only) and that seemed to go well except for the above
problem.
BUT there is still a problem with turn around of bugs/fixes meaning that
zmap is simply not being used enough.
edgrif suggests that he and Roy sit with annotators this time to record bugs and
then immediately go and fix them.
- Navigator with clickable locus names and a display of the clone assembly
is high priority. We need to allow the user to search for locus names within
the navigator window. The locus names will also be displayed in the main
column.
Roy pretty much has this finished now, will be in next release.
- bumping of homols: Matches from the same est will have a background colour
painted between them to make it obvious which matches go with which and perfect
alignments will have "intron" connectors drawn between them. The intron conector
code still needs to be written. This is a high priority item.
edgrif raised performance as a problem when drawing _all_ homol matches as gapped.
Problem is that zmap can end up drawing tens/hundreds of thousands of boxes, most
of which are not needed.
After some discussion it was agreed that code should sift out the "perfect" matches
as defined by a "slop" factor set in the feature style and only draw those as
gapped. They will need to be visually easily distinguishable by the annotators if
this is to work.
edgrif agreed to do a prototype within 2 working days and demo to Jen/Kerstin/others.
- the release notes for each release and the code version numbers are now
available via the help menu in zmap.
They include:
- ZMap build version/release date
- ZMap Request Tracker tickets resolved
- acedb Request Tracker tickets resolved
- summary of ZMap Changes/Fixes
- summary of acedb Changes/Fixes
this enables annotators to see what's been fixed and what code version they are
running at a glance.
- alternative translations: edgrif about half way through code to do this.
- multiple alignments: edgrif is about a third of the way through implementing
a more general way of displaying arbitrary blocks. This is a high priority item.
- zmap needs to get sequences that are in acedb from there for blixem, means
issuing a call to the server to get them but this is possible. NOTE that some
sequences are _only_ in acedb.
- we need a "back" button
- DNA searching in zmap is done.
- There is still a problem with the shutting down of the ace server which
leads to a broken pipe message. edgrif will check that server shuts down
in a timely way.
edgrif said he had trouble reproducing but James has added new information to
request tracker ticket and both jla1 and kj2 said it seemed to be linked to
long running sessions. edgrif will investigate further.
Builds
------
- we can now debug zmap using XCode environment + build universal binaries.
- build system is still a problem as systems are dragging their feet about
nfs mounts on mac boxes, edgrif will have another go and then th will
help expedite.
Blixem/pfetch
--------------
- blixem: dna searching is NOT DONE, edgrif to expedite.
- pfetch: caching? NOT DONE + where should we cache ?
multiple species
----------------
- James and Ed need to discuss how this will be done. Zmap can display
multiple sequences in one zmap window but there will need to changes to
the supporting infrastructure including:
- should the species go in one or multiple acedb databases
- the lace <-> zmap protocol needs to allow specification of
multiple sequence display.
- lace needs to allow the user to pick multiple species
jla1 said a better test than Charlie and NOD mouse is Richard and haplotypes
of which he may have up to 6 (!). jla1 said is there any limit to number, edgrif
said "no, only machine memory....". To quickly test this edgrif will get data from
Richard and try it out to look at performance. Simplest way is to just display
data as is, this mimics what Richard does currently in fmap. edgrif said that
if we used Compara alignment data we could just display the "sub-blocks" of the
alignments that actually contained data. This would produce much more compact
displays.
Acedb
-----
- Treeview "preserve" bug is fixed.
2) Other matters
================
sorry, this section is rather sparse...give me more information if you need to.
- das - NOT DONE (th)
- interface for tags: this is now done thanks to Leo/James.
- submit button for EMBL dumping: needs testing by Jen.
kj2 raised a question about how alignments were done for otter, jgrg said using
BLAST and est2genome. Kerstin said that much more useful alignments could be made
using BLAT or exonerate or ??? In particular, other alignment methods join up their
matches where they are obviously "perfect" overall alignments (e.g. to consecutive
exons) and cope with alignments where they go over clone boundaries.
jgrg to think about this.
==============================================================================
ZMAP_LACE_PROJECT/zmap_lace.2006_11_29 deleted 100755 → 0
+ 0
− 310
View file @ 7c9c2c5a
==============================================================================
ZMap/Otterlace Development
Date: Weds 29th November 2006
Attendees: jla1, jgrg, kj2, edgrif
1) otterlace + zmap progress
============================
new features
------------
- new release: We've sat with Liz/Gavin and that seemed to work really well.
Gavin gave us a list of requests and said that if we could fix the main ones
he would now rather work with zmap than with xace (!).
His list of fixes and status of them are given below, he said that none of them
are show stoppers:
Higher priority:
1/ Showing the co-ordinates/position when you highlight a piece of evidence
2/ Bumping; being able to designate a region based on object/evidence/gene
prediction/position so that everything within or partly within that region will be
clustered; one column per piece of evidence - like xace. Also to have some sort of
shading or connection between pieces of evidence that splice together.
3/ Zooming; being able to designate a region to zoom to, at the moment it's difficult
to zoom out to a particular region in a single click.
4/ CDS not showing up - fixed (?)
5/ Blastx - needs to be strand specific
ALL THESE ARE FIXED.
Medium priority;
6/ 3 frame translation; proteins aren't split up into the 3 separate frames. GF
details needn't be shown automatically in the 3 frames.
Items 6, 8, and 13 below are all the same bug which edgrif will fix.
7/ DNA not highlighted when u click on a piece of evidence or object
8/ Genefinder shouldn't be on as standard.
FIXED
9/ - ve's on the navigator bar when u rev. complement - fixed (?)
FIXED
10/ DNA finder - not visible on the zmap. Would also be nice to be colour coded
depending on the strand. Also it shows the wrong co-ordinates when searching for seq
when you're rev. complemented.
11/ Differentiation between swissprot and trembl - separate columns or different
shading within that column (?)
This is a data issue, we could easily colour these entries differently if that was
what was wanted by giving them seperate methods. We'll do this when edgrif/rds
have written the new style code to replace acedb methods.
Lower priority;
12/ when you click on a locus in the navigator bar for it to zoom to the full locus
rather than the start, (depending on how far you've zoomed in)
FIXED
13/ DNA should be after the evidence rather than before it - actually a funny one -
the first time u turn DNA on it goes between the yellow DNA bar and the
genomic_canonical bar but if u turn it off then on again it goes at the right of the
screen where u want it (?)
14/ Colours when u highlight evidence, esp confusing as hightlighted protein hits look
like cDNA's and vice versa at the moment. Shading like xace has at the moment for
highlighted evidence would be nice.
15/ Does the red band - Genomic_canonical have to be on as standard (?)
FIXED
some additional items from Gavin/Liz:
- can zmap start up "maximised" vertically
FIXED
- can there be a short cut to zoom to a feature
FIXED - new "z" and "Z" short cuts for this. jgrg suggested that these
short cuts should be given on menus/mouse-over popups to remind the
user that they exist...a good idea.
- can the coords for evidence be reported
FIXED
- bumping, can it be still further improved
FIXED edgrif has added a new method for bumping which shows only matches that overlap a transcript.
Works very well, much quicker and much easier to see stuff.
- cursor keys - can now cursor through cols and but for a small glitch could cursor
up and down features in a column.
- bumping of homols: Matches from the same est will have a background colour
painted between them to make it obvious which matches go with which and perfect
alignments will have "intron" connectors drawn between them. The intron conector
code still needs to be written. This is a high priority item.
edgrif has written code to add colour coded bars between the all the hits for a
single match sequence.
The colours are:
green = perfectly colinear
orange = colinear but with missing sequence
red = not colinear
The threshold for a "perfect" match is set via a new tag, "Join_aligns" in the
method for the feature.
There was some discussion about colours, whether we should be using "intron" like
connectors and other aspects but it was agreed to do some trials with users and
get feedback.
edgrif has also added code (following a discussion with annotators) to allow bumping
of _just_ those matches that overlap a selected feature (e.g. transcript). This,
combined with the new zoom function, makes for much more efficient zooming/bumping.
This code needs to be extended to allow the user to select a region and do the bumping
just for that region, this is needed for when there is no existing transcript and the
annotator will make one for that region based on the evidence.
- edgrif raised performance as a problem when drawing _all_ homol matches as gapped.
Problem is that zmap can end up drawing tens/hundreds of thousands of boxes, most
of which are not needed.
After some discussion it was agreed that code should sift out the "perfect" matches
as defined by a "slop" factor set in the feature style and only draw those as
gapped. They will need to be visually easily distinguishable by the annotators if
this is to work.
edgrif agreed to do a prototype within 2 working days and demo to Jen/Kerstin/others.
FIXED
- alternative translations: edgrif about half way through code to do this.
- multiple alignments: edgrif is about a third of the way through implementing
a more general way of displaying arbitrary blocks. This is a high priority item.
- zmap needs to get sequences that are in acedb from there for blixem, means
issuing a call to the server to get them but this is possible. NOTE that some
sequences are _only_ in acedb.
- we need a "back" button, the new zoom functions help with this though.
- There is still a problem with the shutting down of the ace server which
leads to a broken pipe message. edgrif will check that server shuts down
in a timely way.
edgrif said he had trouble reproducing but James has added new information to
request tracker ticket and both jla1 and kj2 said it seemed to be linked to
long running sessions. edgrif will investigate further.
- kj2 would like more information available for features such as the species
and the DE lines. edgrif said they could display this via the current mechanism
of mouse-over popups in the "Info" line at the top of the ZMap. jgrg is going
to make sure this information is in the acedb objects so it can be exported
to zmap.
Builds
------
- build system is still a problem as systems are dragging their feet about
nfs mounts on mac boxes, edgrif will have another go and then th will
help expedite.
FIXED NOW ??
- next build: jgrg has a few features still to add to lace to allow xace to
be removed completely (EUCOMM & other genomic feature editting). He will do
this asap so that we can do a build next Tues to allow users to test on Wens
and Thurs before various people disappear for Christmas. edgrif said they
are ready with zmap and will do a trial build this week for annotators to
try out.
- kj2 said she would like to be able to select a contiguous set of hits and
use them to create a new transcript (much like the "Create Temp Gene" function
in acedb Gene Finder. edgrif and jgrg will look at the code required to do
this.
Blixem/pfetch
--------------
- blixem: dna searching is NOT DONE, edgrif to expedite.
multiple species
----------------
- James and Ed need to discuss how this will be done. Zmap can display
multiple sequences in one zmap window but there will need to changes to
the supporting infrastructure including:
- should the species go in one or multiple acedb databases
- the lace <-> zmap protocol needs to allow specification of
multiple sequence display.
- lace needs to allow the user to pick multiple species
jla1 said a better test than Charlie and NOD mouse is Richard and haplotypes
of which he may have up to 6 (!). jla1 said is there any limit to number, edgrif
said "no, only machine memory....". To quickly test this edgrif will get data from
Richard and try it out to look at performance. Simplest way is to just display
data as is, this mimics what Richard does currently in fmap. edgrif said that
if we used Compara alignment data we could just display the "sub-blocks" of the
alignments that actually contained data. This would produce much more compact
displays.
HAD A MEETING AND ARE IMPLEMENTING CODE IN LACE AND ZMAP TO DO THIS.
Acedb
-----
- Server bug still needs fixing.
2) Other matters
================
sorry, this section is rather sparse...give me more information if you need to.
- script for EMBL dumping: needs testing by Jen.
kj2 raised a question about how alignments were done for otter, jgrg said using
BLAST and est2genome. Kerstin said that much more useful alignments could be made
using BLAT or exonerate or ??? In particular, other alignment methods join up their
matches where they are obviously "perfect" overall alignments (e.g. to consecutive
exons) and cope with alignments where they go over clone boundaries.
jgrg to think about this.
3) Next Meeting
===============
Will be at 11am, 7/12/2006. This will be the last meeting with all of use here
before Christmas.
==============================================================================
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment