search

chr1swallace / coloc

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

Using coloc for conditionally independent eQTLs in the HLA region #113

Open joycekang opened 1 year ago

joycekang commented 1 year ago

Hi Chris,

Thanks for making this fantastic package!

I had a conceptual question - I performed an eQTL analysis within the HLA region, and also did multiple rounds of conditional analysis to find additional independent signals (conditioning on the top variant in each round(s) before). I also did this in 3 cell types. So for example for each gene, I have potentially multiple conditionally independent eQTLs in 3 cell types.

What I'd like to do is compare the eQTLs across cell types to see if any overlap, for example maybe for a given gene, the secondary eQTL signal in cell type 1 colocalizes with the tertiary eQTL signal in cell type 3. Do you think coloc.abf would be appropriate since at each round, I've residualized out the effect of previous round(s)?

Thanks for your advice! Joyce

chr1swallace commented 1 year ago

It could work, if​ you trust your conditional analysis (is this with summary data or original genotypes?)

You could also look at hypercoloc to group the multiple signals

-- https://chr1swallace.github.io

From: Joyce Kang @.> Sent: 13 January 2023 16:44 To: chr1swallace/coloc @.> Cc: Subscribed @.***> Subject: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113)

Hi Chris,

Thanks for making this fantastic package!

I had a conceptual question - I performed an eQTL analysis within the HLA region, and also did multiple rounds of conditional analysis to find additional independent signals (conditioning on the top variant in each round(s) before). I also did this in 3 cell types. So for example for each gene, I have potentially multiple conditionally independent eQTLs in 3 cell types.

What I'd like to do is compare the eQTLs across cell types to see if any overlap, for example maybe for a given gene, the secondary eQTL signal in cell type 1 colocalizes with the tertiary eQTL signal in cell type 3. Do you think coloc.abf would be appropriate since at each round, I've residualized out the effect of previous round(s)?

Thanks for your advice! Joyce

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F113&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZZ44CGbQKRdGauvCSWavZMOobTnpBKBs%2BLwB8%2B7vZ%2F8%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2E5JOA7RZO5NAWSAJTWSGA5ZANCNFSM6AAAAAAT2THTMA&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=4VtJcdIR2G0KvjShY1s1v%2FORNzKKZNd46fCVIfIZD7Y%3D&reserved=0. You are receiving this because you are subscribed to this thread.Message ID: @.***>

joycekang commented 1 year ago

Hi Chris, thanks for the response.

I have summary data (betas, varbetas, pvalues, LD if needed) as well as original genotypes. Is there a better way to do this that needs the original genotypes?

Thanks for the reference to hypercoloc – I’ll check it out!

Joyce

From: Chris Wallace @.> Date: Friday, January 13, 2023 at 11:59 AM To: chr1swallace/coloc @.> Cc: Kang, Joyce @.>, Author @.> Subject: Re: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113) It could work, if​ you trust your conditional analysis (is this with summary data or original genotypes?)

You could also look at hypercoloc to group the multiple signals

-- https://chr1swallace.github.io

From: Joyce Kang @.> Sent: 13 January 2023 16:44 To: chr1swallace/coloc @.> Cc: Subscribed @.***> Subject: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113)

Hi Chris,

Thanks for making this fantastic package!

I had a conceptual question - I performed an eQTL analysis within the HLA region, and also did multiple rounds of conditional analysis to find additional independent signals (conditioning on the top variant in each round(s) before). I also did this in 3 cell types. So for example for each gene, I have potentially multiple conditionally independent eQTLs in 3 cell types.

What I'd like to do is compare the eQTLs across cell types to see if any overlap, for example maybe for a given gene, the secondary eQTL signal in cell type 1 colocalizes with the tertiary eQTL signal in cell type 3. Do you think coloc.abf would be appropriate since at each round, I've residualized out the effect of previous round(s)?

Thanks for your advice! Joyce

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F113&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZZ44CGbQKRdGauvCSWavZMOobTnpBKBs%2BLwB8%2B7vZ%2F8%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2E5JOA7RZO5NAWSAJTWSGA5ZANCNFSM6AAAAAAT2THTMA&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=4VtJcdIR2G0KvjShY1s1v%2FORNzKKZNd46fCVIfIZD7Y%3D&reserved=0. You are receiving this because you are subscribed to this thread.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_chr1swallace_coloc_issues_113-23issuecomment-2D1382136863&d=DwMFaQ&c=WO-RGvefibhHBZq3fL85hQ&r=gkEq751KlSRJlVuy8SraU3Wb28v5yiP_3zek87VnAks&m=CRJuFxgPuhzOzqj9btC1hD0hsj7dD7DORJTTFkqx3M5kVzMXjyvFeNTIJcMXen9R&s=6axcmAiZcSfXd_EPEG-mxfKAY8Nft9bd9-kziaApTGg&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_ADCTGI7EBLXMR4TCCJOMAGDWSGCXLANCNFSM6AAAAAAT2THTMA&d=DwMFaQ&c=WO-RGvefibhHBZq3fL85hQ&r=gkEq751KlSRJlVuy8SraU3Wb28v5yiP_3zek87VnAks&m=CRJuFxgPuhzOzqj9btC1hD0hsj7dD7DORJTTFkqx3M5kVzMXjyvFeNTIJcMXen9R&s=P_CyHheAjEQ7-Qn0vp60bmlhyME6UiYJti9-Ad2cxoA&e=. You are receiving this because you authored the thread.Message ID: @.***>

chr1swallace commented 1 year ago

If you have original genotypes I would run Susie directly on that, particularly in the mhc. I don't thinki have a function yet to work with Susie output but it wouldn't be hard to write one

https://chr1swallace.github.io

From: Joyce Kang @.> Sent: Friday, January 13, 2023 6:38:11 PM To: chr1swallace/coloc @.> Cc: Chris Wallace @.>; Comment @.> Subject: Re: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113)

Hi Chris, thanks for the response.

I have summary data (betas, varbetas, pvalues, LD if needed) as well as original genotypes. Is there a better way to do this that needs the original genotypes?

Thanks for the reference to hypercoloc – I’ll check it out!

Joyce

From: Chris Wallace @.> Date: Friday, January 13, 2023 at 11:59 AM To: chr1swallace/coloc @.> Cc: Kang, Joyce @.>, Author @.> Subject: Re: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113) It could work, if​ you trust your conditional analysis (is this with summary data or original genotypes?)

You could also look at hypercoloc to group the multiple signals

-- https://chr1swallace.github.io

From: Joyce Kang @.> Sent: 13 January 2023 16:44 To: chr1swallace/coloc @.> Cc: Subscribed @.***> Subject: [chr1swallace/coloc] Using coloc for conditionally independent eQTLs in the HLA region (Issue #113)

Hi Chris,

Thanks for making this fantastic package!

I had a conceptual question - I performed an eQTL analysis within the HLA region, and also did multiple rounds of conditional analysis to find additional independent signals (conditioning on the top variant in each round(s) before). I also did this in 3 cell types. So for example for each gene, I have potentially multiple conditionally independent eQTLs in 3 cell types.

What I'd like to do is compare the eQTLs across cell types to see if any overlap, for example maybe for a given gene, the secondary eQTL signal in cell type 1 colocalizes with the tertiary eQTL signal in cell type 3. Do you think coloc.abf would be appropriate since at each round, I've residualized out the effect of previous round(s)?

Thanks for your advice! Joyce

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F113&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=ZZ44CGbQKRdGauvCSWavZMOobTnpBKBs%2BLwB8%2B7vZ%2F8%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2E5JOA7RZO5NAWSAJTWSGA5ZANCNFSM6AAAAAAT2THTMA&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C10f1aad2725f476d48aa08daf5856807%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092250569495571%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=4VtJcdIR2G0KvjShY1s1v%2FORNzKKZNd46fCVIfIZD7Y%3D&reserved=0. You are receiving this because you are subscribed to this thread.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_chr1swallace_coloc_issues_113-23issuecomment-2D1382136863&d=DwMFaQ&c=WO-RGvefibhHBZq3fL85hQ&r=gkEq751KlSRJlVuy8SraU3Wb28v5yiP_3zek87VnAks&m=CRJuFxgPuhzOzqj9btC1hD0hsj7dD7DORJTTFkqx3M5kVzMXjyvFeNTIJcMXen9R&s=6axcmAiZcSfXd_EPEG-mxfKAY8Nft9bd9-kziaApTGg&e=, or unsubscribehttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_ADCTGI7EBLXMR4TCCJOMAGDWSGCXLANCNFSM6AAAAAAT2THTMA&d=DwMFaQ&c=WO-RGvefibhHBZq3fL85hQ&r=gkEq751KlSRJlVuy8SraU3Wb28v5yiP_3zek87VnAks&m=CRJuFxgPuhzOzqj9btC1hD0hsj7dD7DORJTTFkqx3M5kVzMXjyvFeNTIJcMXen9R&s=P_CyHheAjEQ7-Qn0vp60bmlhyME6UiYJti9-Ad2cxoA&e=. You are receiving this because you authored the thread.Message ID: @.***>

— Reply to this email directly, view it on GitHubhttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fchr1swallace%2Fcoloc%2Fissues%2F113%23issuecomment-1382236225&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C607bfbc45f52418f349508daf595660b%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092319256741298%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=%2BA3IIwYc8zt2hyLK5bhaL6rGOAtrOWASjD%2Bxy9jBE4k%3D&reserved=0, or unsubscribehttps://eur03.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAAQWR2EFWOAHG5TCS5BNAT3WSGOJHANCNFSM6AAAAAAT2THTMA&data=05%7C01%7Ccew54%40universityofcambridgecloud.onmicrosoft.com%7C607bfbc45f52418f349508daf595660b%7C49a50445bdfa4b79ade3547b4f3986e9%7C1%7C0%7C638092319256741298%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=%2FA5n5PQuiG37aAD8LBeH05q6rxsQWi7pF4QZ6UMtfdU%3D&reserved=0. You are receiving this because you commented.Message ID: @.***>