docs(roadmap): record where the port stands, measured - #255
Open
BenjaminDEMAILLE wants to merge 1 commit into
Open
docs(roadmap): record where the port stands, measured#255BenjaminDEMAILLE wants to merge 1 commit into
BenjaminDEMAILLE wants to merge 1 commit into
Conversation
The roadmap tracked phases; it did not say how close the output actually is, and the figures quoted elsewhere came from runs nobody could reproduce. Two commands now produce these numbers, and both are in the tree: test/yeast_tier.sh and test/nfcore_diff.py. Yeast tier, 50 000 pairs: 98.658% of mates at the same chromosome, position and CIGAR, 99.964% with the same NH. Of the 1 127 differences, 1 113 are multi-mappers whose primary differs, and 1 110 of those have an identical set of loci on both sides with an identical primary score, which makes them the documented tie-break divergence rather than a gap. That leaves about 15 mates in 84 006, all of them window-formation or stitch-preference differences. Each is named with its CIGAR on both sides, and for the largest one the cause is identified: rerunning it with --alignIntronMax 1000000 reproduces STAR's alignment exactly, so the binning that follows from alignIntronMax=0 is what keeps the two pieces apart.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Speaks to #163 and #143.
The roadmap tracked phases. It did not say how close the output actually is, and the faithfulness figures quoted elsewhere in the tree came from runs nobody could reproduce. Two commands in the tree now produce these numbers:
test/yeast_tier.sh(#253) andtest/nfcore_diff.py(#249).Yeast tier, 50 000 pairs, 84 006 mates
Of the 1 127 mates placed differently, 1 113 are multi-mappers whose primary differs. Those are ties, not disagreements: 1 110 have an identical set of loci on both sides and 1 108 have an identical primary score. That is
DIVERGENCE.md§1.1, already signed off.About 15 mates in 84 006 remain, and the roadmap now names them individually with the CIGAR on each side. The largest group is "STAR splices where we soft-clip", and its cause is identified rather than guessed: re-running
ERR12389696.13842with--alignIntronMax 1000000reproduces STAR's143M831334N7Mexactly, so the binning that follows fromalignIntronMax=0is what keeps the two pieces in windows too far apart to stitch. Scoring is not the problem; window formation is.nf-core/rnaseq, 50 000 pairs
Both gaps users reported on this dataset are fixed (#247 for the buckets, #254 for the depth); the table is what is left after them.
Why this is worth a file rather than a comment
"Finish the port" has been a list of themes. This turns it into a list with numbers next to it: ten themes with pull requests waiting, 19 parameters each owned by one of those pull requests, and ~15 mates whose difference is window formation rather than a missing feature.
🤖 Generated with Claude Code