Codes using artificial neural networks to analyze the binding of coronavirus peptides with polymorphic human MHC class I molecules.
Copyright (C) 2020 Caterina A. M. La Porta and Stefano Zapperi, Center for Complexity and Biosystems, University of Milan
These codes were used to generate results and figures for the paper:
Immune recognition of coronavirus peptides across human populations by artificial neural networks
Caterina A. M. La Porta and Stefano Zapperi
Cell Systems, in press 2020
Codes:
Covid19-4_MHCflurry-chopped-all_alleles.ipynb: This codes runs MHCflurry to obtain predictions for SARS-CoV-2 peptides. Results are filtered with netchop and then saved for further analysis
MHCpan-flurry-covid.ipynb: This code is used to load the csv files containing the predictions for peptides binding affinities obtained with netMHCflurry and netMHCpan4.0 and then filter the results comparing the two methods. The present code is applied to SARS-CoV-2 data.
Effect-of-cutoff.ipynb:This code is used to compare the results for three virus SARS-CoV-2 obtained with a cutoff of 500nM or 1000nM for strong binding
MHCpan-flurry-sars.ipynb: This code is used to load the csv files containing the predictions for peptides binding affinities obtained with netMHCflurry and netMHCpan4.0 and then filter the results comparing the two methods. The present code is applied to SARS-CoV data.
Haplotypes.ipynb: This code allows to compute statistics over populations by distinguishing strongly binding and weakly binding alleles compare-sars-covid-OC43.ipynb: This code is used to compare the results for three virus SARS-CoV-2,SARS-CoV, HCoV-OC43. It also creates various comparative plots
MHCpan-flurry-HCoV-OC43.ipynb: This code is used to load the csv files containing the predictions for peptides binding affinities obtained with netMHCflurry and netMHCpan4.0 and then filter the results comparing the two methods. The present code is applied to HCoV-OC43 data.
netMHCpan-load.ipynb: This code loads predictions from the netMHCpan server, applies a filter from netchop and saves data for further analysis.
TcellPropensity-COVID.ipynb: This code is used to analyzed data from NetTepi
Data:
Data needed to run the codes, including predictions obtained from NetMHCpan and netchop. We also enclose the fasta files for Sars-Cov-2, Sars-Cov and HCoV-OC43