algorithms_bioinformatics

• read-in data

  • visualize the distribution of data to decide which threshold of IC50 to use
  • filter out non-binders (ic50 = 10,000 nM)
  • decide which alleles to use and split the peptide sequences(80/20) for train/test
  • cluster peptides with Hobohm and heuristics (1/r*s)

• construct the PSSM for X MHC I alleles (X = 20)

• visualize the PSSMs as logos

• do prediction with PSSM with test data

• Pearson correlations

The code can be run all by downloading the entire repository and running the __main__.py file. Be aware that in all the files there is a specified data directory which needs to be changed to the new directory where you downloaded the repository. PS: It needs to be fixed so you don't actually need to touch anything from the code and you can call several different variables.