Sprint 16

[cristinaetrv]

Due August 29th

Goal: Writing grant, writing documentation, writing papers on methods, get ML demos (like domain adaptation) out to Thomas's lab to use/try

For IBDGC demo (August 26)

• Build a tool like APOE genotype relevant to IBD and get sum stats out that looks like sample allele counts tool and combined with clinically relevant (one table that looks like APOE chart minus E2/E4 genos, and one table that looks like sample allele counts)

ML

• Make jupyter notebook demos for domain adaptation, adversarial learning, any remaining methods that are undocumented
• Send draft of the paper to David SPPCA to be posted on biorxiv 8/16/24
• Send Mike E POE paper draft 8/9/24
• Update bootstrap package for multi-SNP
• Update permutation test to be less computationally expensive
• Xiucai will give us list of methods, and then we will outline what we want the package to look like
• For Coral/DA method, add geodesics, add tests, generally clean up for use in Thomas's lab

Grant

Backlog:

• Convert dosage matrix to chrom, pos, ref, alt columns instead of locus

• * during PRS calculation, explain PRS, make it fun