Locus breaker in fine-mapping

[dariushghasemi]

We were recommended to use locus breaker working with MLOG10P in COJO pipe rather than the original one.

[dariushghasemi]

Results of running the pipeline for seq.3054.3 (p3 = 1e-4, p4 =1.26e-11):

• 28 independent SNPs identified by COJO
• Solene's filter for jumpers, filter(mlog10p > - log10(p3)), removed one of them.
• Applying locus breaker to reset the boundaries and keeping the variants in the conditional model falling in the new boundaries merely kept 6 SNPs in the end.

Comparing 6 independent SNPs with the standard COJO output (the previous run using p3 =1.26e-11 resulted 6 independent SNPs), only 2 of the 6 SNPs were in common.