staar: GRM parity tests against R ground truth#132
Merged
Conversation
R reimplements the upstream math on a 20-sample, 100-variant fixture and Rust tests check allele freq, G*v, G'*v, PCA, kinship, and unrelated selection against the JSON. Fixes three math bugs the port shipped with: pca.rs squared the T=G'G*U singular values a second time (they're already Σ²) so eigenvalues came out as Σ^4 and related-sample projection divided by d^2 instead of d. Power iteration never kept the Krylov subspace, so score correlations on small cohorts sat around 0.94; replaced with block Krylov accumulation matching drpca.R. unrelated.rs tie-broke divergence backwards: max_by picked the highest, selectUnrel removes the lowest. cargo test: 326/326.
vineetver
force-pushed
the
staar/grm-validation
branch
from
5ab5af1 to
0bd7827
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
R reimplements the upstream math on a 20-sample, 100-variant fixture. Rust tests compare allele freq, G*v, G'*v, PCA, kinship, and unrelated selection against the JSON.
Fixes three math bugs: pca.rs squared T=G'G*U singular values twice so eigenvalues came out as Σ^4 and related-sample projection divided by d^2 instead of d. Power iteration dropped the Krylov subspace so score correlations on small cohorts sat at 0.94; replaced with block Krylov accumulation matching drpca.R. unrelated.rs divergence tie-break was reversed.
cargo test: 326/326.