FeatureDCA

Package used for the analysis of the feature-conditioned autoregressive Direct Coupling Analysis method at:

1. Caredda F., Gennai L., De Los Rios P., Pagnani A., Controllable protein design via autoregressive direct coupling analysis conditioned on principal components, Plos Computational Biology

Install

This is an unregistered package: to install enter ] in the Julia repl and

pkg> add https://github.com/caredda/FeatureDCA.jl

Use

The functions for the training and sampling are

trainer, sample

These take as as inputs either tuples with integer-encoded MSA, weight vector, and feature matrix $(Z,W,Y)$ or a path to the fasta file containing the sequences of the protein family under study along with the feature vector Y. To get more details on the use of each single function use the help function in the Julia repl, e.g.

?trainer

Data

All data used in this study is publicly available at GitHub/francescocaredda/FeatureDCAData in the "data" folder.

Inquiries

Any question can be directed to francesco.caredda@polito.it