How to　reproduce the results in the paper about pypef HOT 5 CLOSED

Thanks a lot :)

from pypef.

from pypef.

This is the result saved on PLS, I replaced some of the evaluation metrics, but did not make any changes to the calculation of R2. From the preserved results, R2 and MSE are seriously inconsistent with the original paper.

from pypef.

Hi,
to reproduce the results, you have to create the training and testing files according to the described split of variant-fitness data. To reproduce results for Dataset B, you can simply run
%run ../../pypef/main.py ml -e aaidx -l LS_B.fasl -t TS_B.fasl --regressor pls as LS_B.fasl and TS_B.fasl are provided in the ANEH folder as example data.

from pypef.

By the way, the used dataset files A-D for PyPEF are available in the SI files of the manuscript (https://ndownloader.figstatic.com/collections/5512280/versions/1). More datasets are e.g. available from https://github.com/Protein-Engineering-Framework/MERGE/tree/main/Data/_variant_fitness_wtseq.

from pypef.