wurmlab / afra Goto Github PK

normal scrolling zoom (with scroll-wheel or gestures on ipad/trackpad) should work: users expect it + its extremely handy + we'll need it for tablets

If two maker predictions overlap (even if on different strands), the curation task is much more difficult. We shouldn't show these to junior users.

When we do show them, the browser should probably be zoomed out to show both models (e.g. this didn't happen for the task at Si_gnF.scaffold02694:11748..16889 )

After clicking "Contribute more", previous curation contribution is still visible (if new location is on same scaffold). Data needs to be flushed/loaded from scratch

When you click "Curate" from the dashboard, the loaded jBrowse scrolls down to the bottom. This is confusing to users. Like Webapollo & normal jBrowse, it should load scrolled to the top (i.e. where you cannot go any higher)

Zoom into a webapollo sequence track. It also shows all 6-frame translations. We obviously need this too.

http://icebox.lbl.gov/WebApolloDemo/jbrowse/?loc=Chr1%3A63809281..63809477&tracks=DNA%2CAnnotations%2CGLEAN%2CNCBI%2Cswissprot_exonerate%2Cillumina_reads_4a%2Cest_spliced%2CEnsembl&highlight=

The file models/task.rb#L52 uses a required keyword argument. The required keywords feature only appeared in Ruby 2.1.0, but the docker image bmpvieira/afra-environment uses ruby 2.0, causing the docker build step to fail in my case. Two possile solutions are

1. @bmpvieira can update his Dockerfile or
2. the required keyword use can be removed?

PLEASE!!! There's no reason for this... and every time you do it, a kitten dies somewhere... srsly!!! THINK OF THE KITTENS!!!1

THIS SEEMS TO ONLY WORK WITH MAKER predictions, not Augustus etc.

I'm using boot2docker on Mac OS 10.10b3, and having loading the app:

docker build -t yeban/afra .
docker run -t -i -p 9292:9292 yeban/afra

This all works find, and I can see that the run.sh script execute and the database is created and app.rb is running and responding to my requests, but whenever I load http://localhost:9292 I get a blank page, not the content from www/index.html.

Looks messy at the moment.

"proposed" tags would be unclear.

all in the title :)

only the first sequence in the edit part shows internal nucleotides. (not the second - e.g. an alternate isoform)

when dragging the screen the yellow edit track stays "stuck" where it was.
Ok but currently the other tracks go on top of this and above this. Visually very confusing - other tracks should disappear underneath the yellow.

After creating a track “Search reference sequence.” of 5 aa (GWYLVW), I scrolled up to check other tracks. Then when I scrolled back down, an error message popped up (see screenshot). I then realised I did not tick “aa” in the track creation, so I created a second track ticking “aa” and exactly the same error message appeared. The only way to remove these error messages was to go back to the dashboard and click “Curate”.

When two windows are open (one for details, one for overview), the edits in one aren't seen in the other.

I tried following the installation instructions on my 10.9.1 mac. Some of the prerequisites were easier (thanks to home-brew). I did the following:

brew install ruby-install
ruby-install ruby  
# actually installs 2.1, not 2.0.0 as the documentation says. 

brew install chruby
# Add the following to the ~/.bashrc or ~/.zshrc file:
# source /usr/local/share/chruby/chruby.sh

brew install https://raw.github.com/postmodern/chgems/master/homebrew/chgems.rb
brew install n

open /Applications/Postgres.app

That all went very smoothly. However, rake fails. The full log is here. At least one reason is that it cannot install pg gem. Instead, you need: gem install pg -- --with-pg-config=/Applications/Postgres.app/Contents/MacOS/bin/pg_config

If I try to push/hack through these issues, I later get stuck with this type of error:

/usr/local/Cellar/ruby/2.1.0/lib/ruby/gems/2.1.0/gems/sequel-4.7.0/lib/sequel/adapters/postgres.rb:161:in async_exec': PG::UndefinedTable: ERROR: relation "features" does not exist (Sequel::DatabaseError)
LINE 1: SELECT * FROM "features" LIMIT 1`

BTW, In the wiki, the line ruby -r ./app.rb -e 'App.gff2jbrowse is missing a closing '

A vertical scroll bar present in Webapollo is missing from the jBrowse

I was editing one model and zoomed out, only to see that the gene model had disappeared from the edit track. I selected it again and dragged it on the edit track. It became shaded (see screenshot) and won't be editable. I had to go back to dashboard and click on curate to fully recover the gene model.

The menu that appears when you right click on a feature to extract sequence etc in webapollo is missing

This is applicable only when zoomed in to base pair level. Perhaps highlight the active reading frame with a green border.

The license information is missing.

Rightclick on exon 2 while resizing exon 1 - the resize thing stays blocked:

Even if you change pages, the resize-thing stays blocked onscreen

The solution is likely to kill any currently active editing action if user right-clicks...

After the first click on a gene model you cannot drag it into the edit track. You can only do this after the second. That it counterintuitive.

Times aren't displayed properly in Safari. Easiest solution is to restrict development & beta usage to Chrome (where things do display correctly).

vertical screenspace is badly used. Only very few gene predictions are shown in comparison with a similar webapollo display

pointer was on RNA coverage track and the automatic display of number of reads appeared. Then when scrolling up to see other tracks, the display area (with number of reads) was still visible (see in screenshot at the bottom). It disappeared when pointer was hovering over it.

After spending a few minutes clicking on things the cursor messes up... e.g. stays stuck in "move around mode" or in "zoom in mode"

Currently need: bullet-points or 'function-less' checkboxes.
Anurag: if you propose something based on webapollo tutorial/documentation in a google doc, I can revise

... this may take one minute or two depending on the speed of your internet connection… thank you for your patience”.

need exon edge matching like in WA.

CSS is messed up when editing (resize boxes, arrowheads)

The search option is not user-friendly. In this particular gene (>maker-Si_gnH.scaffold00024-snap-gene-4.24-mRNA-1Si_gnH.scaffold00024:438016..455209(+) genomic 17193bp), I searched [options aa and forward] for the short sequence VELAG. The new track appeared at the bottom of the screen but was left empty (no usual yellow rectangle with an arrow). When zooming out to 435,000 to 460,000 it can be seen (see screenshot 1). But when zooming in on it (click-and-drag), it disappears (see screenshot 2).

when I zoom in, nucleotides often aren't shown correctly in the exons (within edit track) - the spacing is incorrect thus they are left-aligned within each "block" (the equivalent nucleotides are corruptly positioned in the DNA track)

add WA logo if some WA is used. And/or state in readme that some code is from WA.

when editing, non-canonical splice sites fail to light up

if several models exist in the edit track, they jump around (vertically) as when you resize exons (sometimes)

Can do by modifying or swapping CodonTable.js with another once the following are implemented:

• Afra assumes only one possible start codon (ATG). 'Set Longest ORF', 'Set Translation Start', 'Mark non-canonical translation start site' will need to be flexible to multiple start codons.
• Known splice donors and splice acceptors should be a constant in CodonTable.js, and functions need to take that into account.

"contribute more" shifts me 100bp further to the right. If genes are closely, this makes it feel as if it did nothing.
Instead, "contribute more" should send the user to a random gene with same priority