Sample training data doesn't match the README description

[Habush]

Hi,

I found your paper really interesting and thank you for sharing your work.

It looks like the sample training data doesn't contain information described in the README and the code such as phenotype, age, sex and other covariates. Was this by mistake? If so, can you please share a sample training data so that I can use it to run the model locally? Thanks.

[dlsrnsi]

Hi, I'm very sorry for my late reply. The sample training data is for a drug-response prediction model, and I'm building toy example data for a GWAS model. I also changed the input file to a plink binary file and a covariate file. Does it work for your test running?

[Habush]

Hi,

Yeah, I was able to run the model on the drug data. However, I had to make minor fixes for it work such as adding genotype field to DrugResponseModel as it was not defined but accessed. Also the parent class Genotype2PhenotypeTransformer doesn't accept genotype field so I removed it.

[dlsrnsi]

Oh, you're right. thank you for fixing. I also changed the code too!

[Habush]

Thank you for the fix.

One questions if you explored ways to extend this model to GWAS Summary Statistics data. This would be interesting as summary data is more commonly available than raw genotype data that you used in the paper.

(Perhaps this discussion is worthy of it's own issue. If so, please let me know and I'll create it.)