daiqile96 / OTTERS

A powerful TWAS framework leveraging summary-level reference data
21 stars 1 forks source link

Training/Testing R2 #3

Closed m-mews closed 10 months ago

m-mews commented 1 year ago

Thank you for creating a wonderful pipeline! I ran the pipeline and obtained the weights for the individual methods. The testing R2 for the model was used for filtering the TWAS results; however, I'm struggling to find which script you used to calculate the predicted gene expression values using each method. I found the Summary_OTTERS_GReX.R script in the OTTERS_paper Github, but I do not see how you generated the Pred.GReX files. Thank you in advance for your assistance!

daiqile96 commented 1 year ago

Thank you for pointing out this!

I used OTTERS/imputing.py to calculate the predicted gene expression values (see here).

To use imputing.py, you need to provide a list of samples for imputation. An example of this sample list can be found here. The first column is FID and the second column is IID from the PLINK sample information file (.fam file) of the samples that you want to predict. The example of using imputing.py can also be found in the last section of example.

Sorry for the confusion. Let me know if the imputing.py doesn't work for you.