How many traits can be analyzed at the same time.

[harryyiheyang]

Hi StatFunGen,
I am interested in applying your method to analyze colocalized variants across different gene products, integrating pQTL, eQTL, and sQTL data. Given that a single gene can have multiple sQTLs, the number of traits analyzed simultaneously might reach ten or more. Could you advise on the optimal or recommended number of traits to include when using colocboost?
Thank you!
Yihe

[xueweic]

Hi Yihe,

Thanks for your interest in colocboost! To give you the best guidance, could you let me know whether you are working with individual-level genotype and phenotype data or summary statistics for your xQTL analyses?

If you have individual-level data, we have experience applying colocboost to 50-100 traits simultaneuously, so analyzing ten or more traits should work well in your case. If you are working with summary statistics, I would recommend first evaluating potential summary statistic and LD mismatch issues and performing basic QC steps before running colocboost. We have a set of recommended functions for allele-level QC here: pecotmr/allele_qc.R. As a heads-up, we will also be updating our bioinformatics pipleine within this month, which should make this workflow easier.

Feel free to follow up with more details about your data and we're happy to help further!
Best,
Xuewei

[harryyiheyang]

Hi Xuewei,
Thank you very much for your reply. Indeed, I am planning to use ColocBoost to analyze the xQTL summary data of multiple gene products. As all summary data comes from GTEx and all the LD data comes from the UKBB, I don't think there will be a serious issue of QC. I am looking forward to your bioinformatics pipeline.
Thanks
Yihe