[WIP] Test Suit for Editing Functionality

[hargup]

I need to critically review editing functionality and write automated tests.

• [ ] mergeTranscripts and resizeExon and normalizeFeature each has the same code to "compress" the gene model.
  • [ ] extracted it out into a separate function & write unit tests for the same.
  • [ ] write unit tests for merge, resize, and normalize (normalize may be difficult and can skip for later)
• [ ] Copy relevant tests (or at least data) from Bionode
  • [ ] longest ORF
  • [ ] splice site
  • [ ] ??
  • [ ] if copying data, write appropriate tests. if copying tests, ensure it works in Afra's context.
• [ ] Resize and delete exon have an issue if the edit causes a change to translation start (in 2nd exon?)
  • [ ] fix that
  • [ ] write unit tests for resize and delete
• [ ] If setLongestORF does not find a stop codon, it may:
  • [ ] not introduce CDS in a gene model that does not have CDS. This is not desirable, as gene models created from BLAST evidences would rely on setLongestORF to introduce a CDS. (It could be done via Set Translation Start too, but that's not the norm).

---

Write a Comparison Function for Transcripts/Feature

• [x] Should work even if the name, uniqueID of the transcripts are different.
• [x] The subfeature attribute of a feature is an (ordered) array, this makes the (current) implementation of the comparison function return wrong result. The issue can be resolved in two ways: (1) make sure the order of the subfeatures corresponding a feature is unique. (2) Compute equality of subfeatures as sets. We are using the first method now but it is not thoroughly tested.

---

• [x] Get Input Data
  • [x] RefSeq
  • [x] transcript

---

Testing resizeExon

• [x] The functions on which resizeExon depends needs to tested first
  • [x] getWholeCDSCoordinates
  • [x] setORF
  • [x] getCDNA
  • [x] getCDS
  • [x] transcriptToCDNA
  • [x] CDNAToTranscript
  • [x] setCDS
• [x] No merging and Translation Start and Translation Stop doesn't change.
• [x] The Exon is merged with another exon of the transcript & Translation Start and Translation Stop doesn't change.
• [ ] Cases where Translation Start or Translation Stop changes
• [ ] Verify if CDS are correct.
• [ ] right < left

---

Testing mergeTranscripts

• [ ] The functions on which resizeExon depends needs to tested first
  • [x] areOnSameStrand
  • [ ] sortAnnotationsByLocation
  • [ ] createTranscript
• [ ] With CDS
• [ ] Without CDS
• [ ] The transcripts are on different strands

---

Testing normalizeFeature

• [ ] The functions on which resizeExon depends needs to tested first

---

• [x] Figure out and document various coordinate systems used in the program. I created a wiki: https://github.com/yeban/afra/wiki/Co-ordinate-Systems-used-in-Afra

  Optional:

• [ ] When a test fails when the output and the expected transcript doesn't match it becomes hard to see exactly at which point they were not matching. A possible solution is exporting the transcripts as JSON objects and then using some diff tool to render difference on the test result page.
• [ ] Write a small script to automatically update the transcript_data.js in data/transcripts if there are new transcripts in that directory.

[hargup]

@yeban resizeExon strips the input transcript of all the CDSes. Is this a expected behavior?

[yeban]

It adds back the CDS as well. On Fri, 26 Dec 2014 at 7:16 pm Harsh Gupta notifications@github.com wrote:

@yeban https://github.com/yeban resizeExon strips the input transcript of all the CDSes. Is this a expected behavior?

— Reply to this email directly or view it on GitHub https://github.com/yeban/afra/pull/70#issuecomment-68141653.

[hargup]

@yeban I'm somewhat confused about the about the various coordinate systems used in afra. Here are my observations. Can you verify and comment on them.

• JS and most of the other programming languages uses 0 based indexing and half open intervals. i.,e interval(start, end) is exclusive of the end
• I assume that transcript data uses 1 bases indexing and both side closed interval.
• The docstring of `transcript say

Returns a number between 0 and the length of CDNA of the given transcript minus one (inclusive)

which means CDNA coordinate system uses 0 based indexing and both side closed interval.

[hargup]

After some personal discussion with @yeban I have create this wiki entry https://github.com/yeban/afra/wiki/Co-ordinate-Systems-used-in-Afra

[hargup]

See https://github.com/yeban/afra/issues/61#issuecomment-68996156