search

HaohanWang / thePrecisionLasso

implementation for Precision Lasso: accounting for correlations and linear dependencies in high-dimensional genomic data
https://academic.oup.com/bioinformatics/article/35/7/1181/5089232
MIT License
12 stars 4 forks source link

Expected performance unclear from vignette #2

Open t-silvers opened 5 years ago

t-silvers commented 5 years ago

From the publication and vignette here, it's unclear what is the expected performance of the model. For instance, in BasicExample.ipynb your model assigns weight to each explanatory variable. In the publication, the top 100 genes are selected for further analysis (how? top betas? What parameters yielded those results?).

Some clearer explanations and examples on how you would recommend tuning parameters and training the model would be helpful to include in this repo. For scientific purposes, it would also help in reproducing results from the publication.

Thanks

HaohanWang commented 5 years ago

Hi, we selected the top 100 genes based on the absolute values of betas.

In our paper, we noticed that tuning the parameters according to the desired number of SNPs is a good strategy.

This repository only serves as an example for people who are interested to use it, if you want to replicate the results for scientific purpose, I'm happy to share the private repo with all the scripts with you. (I will edit the README to avoid further confusion of others about this point.)