staar: flip coding-mask predicates to MetaSVM#127
Merged
Conversation
STAARpipeline gates disruptive_missense on (nonsynonymous SNV) & (MetaSVM_pred=="D"). We were proxying with CADD>20 OR REVEL>0.5, so the three dependent masks (disruptive_missense, plof_ds, ptv_ds) returned p-values that didn't match R. CADD and REVEL are weight channels in the score kernel, not damage classifiers. disruptive_missense now reads annotation.metasvm_pred directly. is_plof_ds = plof ∪ disruptive_missense. is_ptv_ds = ptv ∪ splice ∪ disruptive_missense (matches ptv_ds.R's variant_type=="variant" branch where splicing is part of the union unconditionally; the old splice+CADD>20 gate was our invention, not R's). PLofMissense renamed to PLofDs; old is_plof_or_missense predicate was the broad plof ∪ all-missense union which R does not ship. The docs-only example output path coding_pLoF_missense.parquet becomes coding_plof_ds.parquet. Test helper v() drops cadd/revel args since no predicate reads them now. New v_ds() helper sets metasvm_pred=Deleterious for disruptive-missense tests. Four old CADD/REVEL-threshold tests deleted; two new tests pin the MetaSVM gate directly. Net: 105 fewer lines, 291 tests green.
vineetver
force-pushed
the
staar/77-coding-mask-metasvm-flip
branch
from
2d62651 to
8df2c9d
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #77. Depends on #126 (A2 MetaSVM column wiring) — merge that first.
STAARpipeline gates disruptive_missense on (nonsynonymous SNV) & (MetaSVM_pred=="D"). We were proxying with CADD>20 OR REVEL>0.5, so disruptive_missense / plof_ds / ptv_ds returned wrong p-values. CADD and REVEL are weight channels in the score kernel, not damage classifiers.
disruptive_missense now reads annotation.metasvm_pred. plof_ds = plof ∪ disruptive_missense. ptv_ds = ptv ∪ splice ∪ disruptive_missense — matches R's ptv_ds.R variant_type="variant" branch where splicing is in the union unconditionally; the old splice+CADD>20 gate was ours, not R's. PLofMissense renamed to PLofDs.
Skipped the score-cache predicate-version bust that the plan mentioned. The cache stores per-variant U and K which are mask-independent; masks select which variants feed into a given gene test at scoring time. Flipping the predicate doesn't invalidate cached U/K.
291/291, clippy clean. Four CADD/REVEL-threshold tests deleted, two MetaSVM tests added; net -105 lines.