search

chr1swallace / coloc

Repo for the R package coloc
144 stars 44 forks source link

How to interpret coloc.susie results with two different variants in low LD? #98

Closed wonjikim11 closed 2 years ago

wonjikim11 commented 2 years ago

Hi,

I visited https://github.com/chr1swallace/coloc/issues/84 and figured out how to interpret coloc.susie results with two different variants in high LD. In most cases, they were in high LD but some were in low LD such as r2=0.5, and I wonder how to interpret this. Can I interpret this as multiple causal variants?

One more, if I found two colocalizations involving one eQTL and three variants, e.g. eQTL1 - variant1 & variant2 eQTL1 - variant1 & variant3, can it be considered one colocalization involving three variants?

Thanks!

chr1swallace commented 2 years ago

I think you need to give me more detail.

Are the two Susie variants in the same credible set or different?

What do you mean by your eqtl colocalisations? Coloc doesn't give you a yes/no answer, but a probability. What do the data look like for these traits (use plot_dataset()) and what method of analysis are you using?

https://chr1swallace.github.io

From: wonjikim11 @.> Sent: Tuesday, August 16, 2022 6:19:07 AM To: chr1swallace/coloc @.> Cc: Subscribed @.***> Subject: [chr1swallace/coloc] How to interpret coloc.susie results with two different variants in low LD? (Issue #98)

Hi,

I visited #84https://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F84&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C65f14bb703f24665ce3e08da7f46eb80%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962239813854208%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=vDrmOXmfQqoHatdUjxIt%2Br3ijQ3B%2F4Y%2F5GGNGrhSaM0%3D&reserved=0 and figured out how to interpret coloc.susie results with two different variants in high LD. In most cases, they were in high LD but some were in low LD such as r2=0.5, and I wonder how to interpret this. Can I interpret this as multiple causal variants?

One more, if I found two colocalizations involving one eQTL and three variants, e.g. eQTL1 - variant1 & variant2 eQTL1 - variant1 & variant3, can it be considered one colocalization involving three variants?

Thanks!

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F98&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C65f14bb703f24665ce3e08da7f46eb80%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962239814010444%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=KV3dUivgxE6LapjsKBJt90J13c7eGTcCVI%2FMuDkfX3E%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2DQJEHLXCJXNUUYSSDVZMQEXANCNFSM56UM25UQ&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C65f14bb703f24665ce3e08da7f46eb80%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962239814010444%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=62CGMtcPAjQUTRN4zI0T8d0HH7Dpm9cd6Pi14yEqcJI%3D&reserved=0. You are receiving this because you are subscribed to this thread.Message ID: @.***>

wonjikim11 commented 2 years ago
  1. They are in the same credible set.
  2. Let me take a look at the plot_dataset(). I used coloc.susie on gwas and eqtl with summary statistics. I wonder if multiple colocalizations sharing a causal variant can be interpreted as a single colocalization with multiple causal variants.
chr1swallace commented 2 years ago

1 if in the same credible set, they are not two signals 2 suspect this relates to (lack of) power to distinguish signals

https://chr1swallace.github.io

From: wonjikim11 @.> Sent: Tuesday, August 16, 2022 7:16:05 AM To: chr1swallace/coloc @.> Cc: Chris Wallace @.>; Comment @.> Subject: Re: [chr1swallace/coloc] How to interpret coloc.susie results with two different variants in low LD? (Issue #98)

  1. They are in the same credible set.
  2. Let me take a look at the plot_dataset(). I used coloc.susie on gwas and eqtl with summary statistics. I wonder if multiple colocalizations sharing a causal variant can be interpreted as a colocalization with multiple causal variants.

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F98%23issuecomment-1216192398&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C4ce6cdc94bf74a58490008da7f4ece53%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962273681955068%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=JZLBhxfek%2BLCfPsYXMVKwg6CqUhDxcM76JM616%2BjWtI%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2BNJGFBC4IZ7PBGN2LVZMW2LANCNFSM56UM25UQ&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C4ce6cdc94bf74a58490008da7f4ece53%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962273681955068%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=figMu4K9IohbVfV317TqjUS8hsUen%2FwCIbU%2BObPM%2FgM%3D&reserved=0. You are receiving this because you commented.Message ID: @.***>

wonjikim11 commented 2 years ago
  1. So, in what cases can we conclude that there are multiple causal variants? If they are in the different credible sets?
chr1swallace commented 2 years ago

yes

From: wonjikim11 @.> Sent: 16 August 2022 15:50 To: chr1swallace/coloc @.> Cc: Chris Wallace @.>; Comment @.> Subject: Re: [chr1swallace/coloc] How to interpret coloc.susie results with two different variants in low LD? (Issue #98)

  1. So, in what cases can we conclude that there are multiple causal variants? If they are in the different credible sets?

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F98%23issuecomment-1216745756&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7Ca4994740c1524a093f2108da7f96b8eb%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962582561443839%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=P7HH41R3rbFeSgExUnuhIOE3txAxO0o%2FdAK4OSEc0c8%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2FQKPDC7JT2TFR4TR3VZOTEZANCNFSM56UM25UQ&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7Ca4994740c1524a093f2108da7f96b8eb%7C49a50445bdfa4b79ade3547b4f3986e9%7C0%7C0%7C637962582561600093%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=RK2XiNGrhfejHXzydbBtQ4NH9f7ThkWg3qBigClkStI%3D&reserved=0. You are receiving this because you commented.Message ID: @.***>

wonjikim11 commented 2 years ago

I got it. Thank you!