Mismatch in some, not all, assoc-agg outputs

[aofarrel]

weights

• chr1: Within acceptable range of tolerance (0.00000001)
• chr2: md5 identical

position:

• chr1: Do not match even with tolerance of 1000
• chr2: Within acceptable range of tolerance (0.00000001)

allele:

• chr1: Within acceptable range of tolerance (0.00000001)
• chr2: Within acceptable range of tolerance (0.00000001)

[aofarrel]

Here's allele's chr1's position files in Rstudio, Terra vs SB. Something fishy is going on. edit jan 2021: I think "allele's chr1's position files" is supposed to be "chr1's position files" but I don't know if it meant assoc-agg the task or assoc-agg the workflow

But their graphs are essentially equivalent across SB and Terra. So even though there is an extra column in one of these, it seems that isn't affecting the plotting step.

Terra QQ

SB QQ

Terra manh

SB manh

[aofarrel]

If we look at weights instead...

assoccwl: weights_rerun_sbg_chr1_seg1.RData as a group_segments task output assocwdl: 1KG_phase3_subset_annot.RData as an assoc_aggregate task input

all.equal(assoccwl, assocwdl, tolerance=1e-100) [1] "Component “results”: Component “Score”: Mean relative difference: 1.48677959e-16"
[2] "Component “results”: Component “Score.Stat”: Mean relative difference: 1.32878924e-16" [3] "Component “results”: Component “Score.pval”: Mean relative difference: 6.739930668e-15" [4] "Component “results”: Component “Est”: Mean relative difference: 2.375174983e-16"
[5] "Component “results”: Component “PVE”: Mean relative difference: 3.367141493e-16"
options(digits=22) print(assocwdl[["results"]]["Score"]) Score 375790 11.94768109113651455289 print(assoccwl[["results"]]["Score"]) Score 375790 11.94768109113651632924

[aofarrel]

This is still an issue but given the big changes that happened upstream, it is worth making a new ticket for it.