Using my own eQTL data

[YingboHuang]

Hi,

I have a question regarding using S-PrediXcan with my own eQTL summary statistics. What would be the first step to go? I see the repository (QTL_to_PredictDB), is that the one I should first go to?

Thanks for your time!

Yingbo

[Heroico]

Hi Tingbo,

Indeed that repository provides one way to generate S-PrediXcan (and PrediXcan) models from QTL summary statistics.
If you have multiple tissues, you can instead ran MASHR on the summary statistics and use the corrected effect sizes to build linear models for experesion.

Best,

Alvaro

[YingboHuang]

Hi Alvaro,

Thank you so much for such a quick response. For the case that I only have eQTL summary statistics from 1 tissue, what is the best way to generate the model? (I assume that MASHR does not work for single tissue here).

Thanks in advance,
Yingbo