Are there alternatives to LOBICO? Logic combinations to predict a response variable.

Question

[brianj.park]

Hi all, I am trying to create a model of drug response vs. mutation data in certain cell lines, where I can study the effect of a combination of the mutations (i.e., combination of feature variables in the model to predict the response). This Nature paper describes essentially what I need: https://www.nature.com/articles/srep36812. However, their program LOBICO seem to be unsupported and the adapted R package RLOBICO (https://github.com/bhklab/RLOBICO) refuses to run on my hardware. Are there alternatives to do something similar? Something like a random forests model could work as well but I don't think I can extract combinatory effects between the predictors.

Thank you.

Answer 1

[aloesch.fach1]

Hi Brijani,

Have you found a solution to your issue so far? I'd be interested as well. Btw, the paper where the LOBICO makers apply their method (https://www.cell.com/fulltext/S0092-8674(16)30746-2) mentions random forest and EN regression models as well, maybe having a look into that could help you.

Best wishes, Andi