LD matrix on SuSIE_coloc
Zhangzzzzzy opened this issue · 7 comments
As a coloc_susie newbie, I have this question, how should I calculate the LD matrix to get it? Or how to apply it to, I didn't find any information about it, looking forward to your answers.
您需要从参考总体中找到与 GWAS 研究中使用的总体匹配的 LD 矩阵。Doug Speed 有一个很好的示例脚本，如何在这里为欧洲人口生成一个脚本 https://dougspeed.com/reference-panel/
...
-- https://chr1swallace.github.iohttps://chr1swallace.github.io/
________________________________ From: Zhangzzzzzy @.> Sent: Friday, September 29, 2023 3:45 AM To: chr1swallace/coloc @.> Cc: Subscribed @.> Subject: [chr1swallace/coloc] LD matrix on SuSIE_coloc (Issue #132) As a coloc_susie newbie, I have this question, how should I calculate the LD matrix to get it? Or how to apply it to, I didn't find any information about it, looking forward to your answers. — Reply to this email directly, view it on GitHub<#132>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAQWR2ERLC5XVDCWDXGT4CLX4YY4VANCNFSM6AAAAAA5L2VN5E. You are receiving this because you are subscribed to this thread.Message ID: @.>
Hello, I am glad to receive such a quick reply from you. After reading the information you shared, I have new questions. First of all, the way you shared the processed file results in three files: .bed, .bim, .fam, which one is the LD matrix I need to use? Second question: I am using GWAS aggregated data from European populations, can I directly use the template from 1000g European populations? How can I apply this template to coloc_SuSiE? I noticed that the LD matrix in the reference data is a matrix with snp number of squares on the sides, but obviously the reference template is not.
doug speed's script should show you how to create the LD matrix from those files
…
-- https://chr1swallace.github.iohttps://chr1swallace.github.io/
________________________________ From: Zhangzzzzzy @.> Sent: Friday, September 29, 2023 10:22 AM To: chr1swallace/coloc @.> Cc: Chris Wallace @.>; Comment @.> Subject: Re: [chr1swallace/coloc] LD matrix on SuSIE_coloc (Issue #132) 您需要从参考总体中找到与 GWAS 研究中使用的总体匹配的 LD 矩阵。Doug Speed 有一个很好的示例脚本，如何在这里为欧洲人口生成一个脚本 https://dougspeed.com/reference-panel/ ... -- https://chr1swallace.github.iohttps://chr1swallace.github.io/https://chr1swallace.github.io/
________________________________ From: Zhangzzzzzy @.> Sent: Friday, September 29, 2023 3:45 AM To: chr1swallace/coloc @.> Cc: Subscribed @.> Subject: [chr1swallace/coloc] LD matrix on SuSIE_coloc (Issue #132<#132>) As a coloc_susie newbie, I have this question, how should I calculate the LD matrix to get it? Or how to apply it to, I didn't find any information about it, looking forward to your answers. ― Reply to this email directly, view it on GitHub<#132<#132>>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAQWR2ERLC5XVDCWDXGT4CLX4YY4VANCNFSM6AAAAAA5L2VN5E. You are receiving this because you are subscribed to this thread.Message ID: @.> Hello, I am glad to receive such a quick reply from you. After reading the information you shared, I have new questions. First of all, the way you shared the processed file results in three files: .bed, .bim, .fam, which one is the LD matrix I need to use? Second question: I am using GWAS aggregated data from European populations, can I directly use the template from 1000g European populations? How can I apply this template to coloc_SuSiE? I noticed that the LD matrix in the reference data is a matrix with snp number of squares on the sides, but obviously the reference template is not. ― Reply to this email directly, view it on GitHub<#132 (comment)>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAQWR2AXDUIH3BZ3OPFDVADX42HNVANCNFSM6AAAAAA5L2VN5E. You are receiving this because you commented.Message ID: @.***>
Okay, thank you very much for your reply!
You need to find an LD matrix from a reference population that matches the population used in your GWAS study. Doug Speed has a nice example script how to generate one for a European population here https://dougspeed.com/reference-panel/
…
Do you need to construct a LD matrix for each of the two datasets involved in susie analysis or just one LD matrix for all SNPs in the two datasets?
BTW, IMHO, it would be much better if coloc.susie can take care of the LD information by itself because "coloc" 5.2.3 has all the relevant (including LD) information.
Thanks a lot.
Does this functions work?
Get LD matrix for list of SNPs
https://mrcieu.github.io/TwoSampleMR/reference/ld_matrix.html
And
Does this work？