How to interpret the results of susie summary？

[Leweibo]

https://chr1swallace.github.io/coloc/articles/a06_SuSiE.html

In the article, the example provided by the author, there are two rows in the result of susie.res$summary.

nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf
1: 500 s105 s105 3.079008e-14 6.507291e-07 1.342030e-10 0.0008379729
2: 500 s89 s105 1.422896e-06 2.209787e-04 6.201896e-03 0.9631063075
PP.H4.abf idx1 idx2
1: 0.99916138 1 1
2: 0.03046939 2 1

Results pass decision rule H4 > 0.9

Results fail decision rule H4 > 0.9

In my study, there are even more rows in the result of susie.res$summary

print(susie.res$summary)
nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2

1: 4591 rs12509595 rs1458038 0.000000e+00 4.147380e-13 0.000000e+00 0.02012341 9.798766e-01 1 1
2: 4591 rs10213506 rs1458038 1.570304e-273 1.878073e-11 8.361250e-263 1.00000000 6.399585e-12 2 1
3: 4591 rs74780855 rs1458038 5.954240e-106 1.878073e-11 3.170398e-95 1.00000000 1.328367e-13 3 1
4: 4591 rs72661739 rs1458038 1.556576e-79 1.878073e-11 8.288157e-69 1.00000000 1.963290e-13 4 1
5: 4591 rs10006582 rs1458038 1.476746e-59 1.878073e-11 7.863090e-49 1.00000000 6.277067e-13 5 1
6: 4591 rs6848130 rs1458038 2.330814e-86 1.878073e-11 1.241066e-75 1.00000000 2.036101e-09 6 1
7: 4591 rs2867702 rs1458038 1.508344e-52 1.878073e-11 8.031337e-42 1.00000000 3.088207e-13 7 1
8: 4591 rs10029510 rs1458038 6.966293e-45 1.878073e-11 3.709276e-34 1.00000000 1.354325e-12 8 1
9: 4591 rs7668598 rs1458038 1.734194e-52 1.878073e-11 9.233902e-42 1.00000000 1.516226e-12 9 1
10: 4591 rs1987331 rs1458038 1.352851e-41 1.878073e-11 7.203400e-31 1.00000000 3.023677e-13 10 1

My results Example 3:

print(susie.res2$summary)
nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2

1: 5296 rs13335818 rs77924615 0.000000e+00 1.278006e-114 0.000000e+00 1.0000000 1.774691e-75 1 1
2: 5296 rs28510439 rs77924615 0.000000e+00 1.278006e-114 5.663628e-318 1.0000000 2.785318e-115 2 1
3: 5296 rs77924615 rs77924615 0.000000e+00 2.556012e-117 1.709512e-286 0.0000000 1.000000e+00 3 1
4: 5296 rs190017805 rs77924615 5.049465e-249 1.278006e-114 3.951049e-135 1.0000000 4.971996e-117 4 1
5: 5296 rs62032857 rs77924615 7.843335e-207 1.278006e-114 6.137165e-93 1.0000000 9.194789e-91 5 1
6: 5296 rs149109606 rs77924615 1.310727e-218 1.278006e-114 1.025603e-104 1.0000000 1.522628e-106 6 1
7: 5296 rs16971906 rs77924615 8.365323e-194 1.278006e-114 6.545605e-80 1.0000000 7.159603e-82 7 1
8: 5296 rs76621572 rs77924615 7.385114e-188 1.278006e-114 5.778622e-74 1.0000000 4.188686e-66 8 1
9: 5296 rs75044573 rs77924615 2.609464e-190 1.278006e-114 2.041824e-76 1.0000000 1.359528e-70 9 1
10: 5296 rs12598673 rs77924615 1.108288e-174 1.278006e-114 8.672005e-61 1.0000000 9.313257e-63 10 1
11: 5296 rs13335818 rs71373185 0.000000e+00 3.373833e-41 0.000000e+00 1.0000000 1.345593e-39 1 3
12: 5296 rs28510439 rs71373185 0.000000e+00 3.373833e-41 5.663628e-318 1.0000000 1.196660e-41 2 3
13: 5296 rs77924615 rs71373185 4.940656e-324 3.373833e-41 8.547561e-284 1.0000000 2.969029e-30 3 3
14: 5296 rs190017805 rs71373185 1.333018e-175 3.373833e-41 3.951049e-135 1.0000000 3.894044e-43 4 3
15: 5296 rs62032857 rs71373185 2.070577e-133 3.373833e-41 6.137165e-93 1.0000000 5.975595e-37 5 3
16: 5296 rs149109606 rs71373185 3.460213e-145 3.373833e-41 1.025603e-104 1.0000000 9.895198e-39 6 3
17: 5296 rs16971906 rs71373185 2.208377e-120 3.373833e-41 6.545605e-80 1.0000000 1.244894e-43 7 3
18: 5296 rs76621572 rs71373185 1.949610e-114 3.373833e-41 5.778622e-74 1.0000000 2.304057e-43 8 3
19: 5296 rs75044573 rs71373185 6.888773e-117 3.373833e-41 2.041824e-76 1.0000000 1.906309e-43 9 3
20: 5296 rs12598673 rs71373185 2.925789e-101 3.373833e-41 8.672005e-61 1.0000000 1.922213e-43 10 3
21: 5296 rs13335818 rs7198770 0.000000e+00 3.448060e-14 0.000000e+00 1.0000000 2.081256e-16 1 5
22: 5296 rs28510439 rs7198770 0.000000e+00 3.448060e-14 5.663628e-318 1.0000000 2.817607e-16 2 5
23: 5296 rs77924615 rs7198770 2.947251e-297 3.448060e-14 8.547561e-284 1.0000000 3.148348e-16 3 5
24: 5296 rs190017805 rs7198770 1.362345e-148 3.448060e-14 3.951049e-135 1.0000000 9.574237e-16 4 5
...
69: 5296 rs75044573 rs4494548 2.506562e-88 1.227609e-12 2.041824e-76 1.0000000 9.663381e-14 9 4
70: 5296 rs12598673 rs4494548 1.064583e-72 1.227609e-12 8.672005e-61 1.0000000 7.188687e-15 10 4
nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2

Question:
How to interpret the results of susie summary? which rows of PP.H4.abf is the Coloc results?

If any of the results (rows) pass decision rule H4 > 0.9, then the results as a whole pass the decision rule?

[chr1swallace]

I just replied to you asking the same question in issue #87. Susie appears to have found 10 distinct signals for trait 1 and 7 for trait 2. coloc then tries to colocalise each pair- 70 tests of colocalisation. Are you sure you have the correct LD matrix - is it from the same population as your GWAS data and are the alleles aligned correctly? Looking at the manhattan plots, is it reasonable that there are 10 and 7 separate signals for each trait?

…

-- https://chr1swallace.github.io<https://chr1swallace.github.io/>

________________________________ From: WB ***@***.***> Sent: Wednesday, March 20, 2024 10:42 AM To: chr1swallace/coloc ***@***.***> Cc: Subscribed ***@***.***> Subject: [chr1swallace/coloc] How to interpret the results of susie summary？ (Issue #151) https://chr1swallace.github.io/coloc/articles/a06_SuSiE.html In the article, the example provided by the author, there are two rows in the result of susie.res$summary. nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf 1: 500 s105 s105 3.079008e-14 6.507291e-07 1.342030e-10 0.0008379729 2: 500 s89 s105 1.422896e-06 2.209787e-04 6.201896e-03 0.9631063075 PP.H4.abf idx1 idx2 1: 0.99916138 1 1 2: 0.03046939 2 1 Results pass decision rule H4 > 0.9 Results fail decision rule H4 > 0.9 In my study, there are even more rows in the result of susie.res$summary print(susie.res$summary) nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2 1: 4591 rs12509595 rs1458038 0.000000e+00 4.147380e-13 0.000000e+00 0.02012341 9.798766e-01 1 1 2: 4591 rs10213506 rs1458038 1.570304e-273 1.878073e-11 8.361250e-263 1.00000000 6.399585e-12 2 1 3: 4591 rs74780855 rs1458038 5.954240e-106 1.878073e-11 3.170398e-95 1.00000000 1.328367e-13 3 1 4: 4591 rs72661739 rs1458038 1.556576e-79 1.878073e-11 8.288157e-69 1.00000000 1.963290e-13 4 1 5: 4591 rs10006582 rs1458038 1.476746e-59 1.878073e-11 7.863090e-49 1.00000000 6.277067e-13 5 1 6: 4591 rs6848130 rs1458038 2.330814e-86 1.878073e-11 1.241066e-75 1.00000000 2.036101e-09 6 1 7: 4591 rs2867702 rs1458038 1.508344e-52 1.878073e-11 8.031337e-42 1.00000000 3.088207e-13 7 1 8: 4591 rs10029510 rs1458038 6.966293e-45 1.878073e-11 3.709276e-34 1.00000000 1.354325e-12 8 1 9: 4591 rs7668598 rs1458038 1.734194e-52 1.878073e-11 9.233902e-42 1.00000000 1.516226e-12 9 1 10: 4591 rs1987331 rs1458038 1.352851e-41 1.878073e-11 7.203400e-31 1.00000000 3.023677e-13 10 1 My results Example 3: print(susie.res2$summary) nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2 1: 5296 rs13335818 rs77924615 0.000000e+00 1.278006e-114 0.000000e+00 1.0000000 1.774691e-75 1 1 2: 5296 rs28510439 rs77924615 0.000000e+00 1.278006e-114 5.663628e-318 1.0000000 2.785318e-115 2 1 3: 5296 rs77924615 rs77924615 0.000000e+00 2.556012e-117 1.709512e-286 0.0000000 1.000000e+00 3 1 4: 5296 rs190017805 rs77924615 5.049465e-249 1.278006e-114 3.951049e-135 1.0000000 4.971996e-117 4 1 5: 5296 rs62032857 rs77924615 7.843335e-207 1.278006e-114 6.137165e-93 1.0000000 9.194789e-91 5 1 6: 5296 rs149109606 rs77924615 1.310727e-218 1.278006e-114 1.025603e-104 1.0000000 1.522628e-106 6 1 7: 5296 rs16971906 rs77924615 8.365323e-194 1.278006e-114 6.545605e-80 1.0000000 7.159603e-82 7 1 8: 5296 rs76621572 rs77924615 7.385114e-188 1.278006e-114 5.778622e-74 1.0000000 4.188686e-66 8 1 9: 5296 rs75044573 rs77924615 2.609464e-190 1.278006e-114 2.041824e-76 1.0000000 1.359528e-70 9 1 10: 5296 rs12598673 rs77924615 1.108288e-174 1.278006e-114 8.672005e-61 1.0000000 9.313257e-63 10 1 11: 5296 rs13335818 rs71373185 0.000000e+00 3.373833e-41 0.000000e+00 1.0000000 1.345593e-39 1 3 12: 5296 rs28510439 rs71373185 0.000000e+00 3.373833e-41 5.663628e-318 1.0000000 1.196660e-41 2 3 13: 5296 rs77924615 rs71373185 4.940656e-324 3.373833e-41 8.547561e-284 1.0000000 2.969029e-30 3 3 14: 5296 rs190017805 rs71373185 1.333018e-175 3.373833e-41 3.951049e-135 1.0000000 3.894044e-43 4 3 15: 5296 rs62032857 rs71373185 2.070577e-133 3.373833e-41 6.137165e-93 1.0000000 5.975595e-37 5 3 16: 5296 rs149109606 rs71373185 3.460213e-145 3.373833e-41 1.025603e-104 1.0000000 9.895198e-39 6 3 17: 5296 rs16971906 rs71373185 2.208377e-120 3.373833e-41 6.545605e-80 1.0000000 1.244894e-43 7 3 18: 5296 rs76621572 rs71373185 1.949610e-114 3.373833e-41 5.778622e-74 1.0000000 2.304057e-43 8 3 19: 5296 rs75044573 rs71373185 6.888773e-117 3.373833e-41 2.041824e-76 1.0000000 1.906309e-43 9 3 20: 5296 rs12598673 rs71373185 2.925789e-101 3.373833e-41 8.672005e-61 1.0000000 1.922213e-43 10 3 21: 5296 rs13335818 rs7198770 0.000000e+00 3.448060e-14 0.000000e+00 1.0000000 2.081256e-16 1 5 22: 5296 rs28510439 rs7198770 0.000000e+00 3.448060e-14 5.663628e-318 1.0000000 2.817607e-16 2 5 23: 5296 rs77924615 rs7198770 2.947251e-297 3.448060e-14 8.547561e-284 1.0000000 3.148348e-16 3 5 24: 5296 rs190017805 rs7198770 1.362345e-148 3.448060e-14 3.951049e-135 1.0000000 9.574237e-16 4 5 ... 69: 5296 rs75044573 rs4494548 2.506562e-88 1.227609e-12 2.041824e-76 1.0000000 9.663381e-14 9 4 70: 5296 rs12598673 rs4494548 1.064583e-72 1.227609e-12 8.672005e-61 1.0000000 7.188687e-15 10 4 nsnps hit1 hit2 PP.H0.abf PP.H1.abf PP.H2.abf PP.H3.abf PP.H4.abf idx1 idx2 Question: How to interpret the results of susie summary? which rows of PP.H4.abf is the Coloc results? If any of the results (rows) pass decision rule H4 > 0.9, then the results as a whole pass the decision rule? — Reply to this email directly, view it on GitHub<#151>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAQWR2B23ZO7M5FITJXTUH3YZFRYBAVCNFSM6AAAAABE7GX26CVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE4TOMRQHA4DMNQ>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Leweibo]

Thanks for responding.

Perhaps I have got the answer here:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9746220/

Under the assumption of only a single causal variant, if there were multiple instrumental variables used in MR, we calculated the average PP.H4 from the coloc.abf output. Under the assumption of multiple causal variants exist [28], we used the maximum PP.H4 of multiple credible sets from the coloc.susie output.

[chr1swallace]

My strong advice is to check that your Susie results make sense. I doubt you have so many casual variants really. I fear your LD matrix is not a good match for your population, either because there's a real mismatch or because there are allele ordering errors. Please do check the Susie results make sense before trying to interpret the coloc results.

…

-- https://chr1swallace.github.io

________________________________ From: WB ***@***.***> Sent: Thursday, March 21, 2024 5:05:28 PM To: chr1swallace/coloc ***@***.***> Cc: Chris Wallace ***@***.***>; Comment ***@***.***> Subject: Re: [chr1swallace/coloc] How to interpret the results of susie summary？ (Issue #151) Thanks for responding. Perhaps I have got the answer here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9746220/ Under the assumption of only a single causal variant, if there were multiple instrumental variables used in MR, we calculated the average PP.H4 from the coloc.abf output. Under the assumption of multiple causal variants exist [28<https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9746220/#CR28>], we used the maximum PP.H4 of multiple credible sets from the coloc.susie output. — Reply to this email directly, view it on GitHub<#151 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAQWR2CXDGJLHGBXIXMDC2TYZMHNRAVCNFSM6AAAAABE7GX26CVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAMJTGA2DCNBRGE>. You are receiving this because you commented.Message ID: ***@***.***>