Update reference proteome analysis to consider patient genotype

[malachig]

Currently we compare peptide sequences against a reference proteome from Ensembl.

It would be more correct to create a proteome that incorporated SNPs that impact protein sequence and then look for matches between our peptide sequence and THAT personalized proteome reference.

We could take our germline VCF, use it to update a reference genome fasta with het/hom SNPs, extract CDS sequences based on all annotated protein coding transcripts from Ensembl and then build the peptide reference for comparison from that.

[malachig]

At least two complexities to this:

• Insertions and deletions
• Diploid representation

[malachig]

For a description of a graph based approach to tackling this problem refer to section: "Derivation of Splicing-Derived Peptides" from Kahles et al. "Comprehensive Analysis of Alternative Splicing Across Tumors from 8,705 Patients."

https://www.sciencedirect.com/science/article/pii/S1535610818303064?via%3Dihub#sec5

[susannasiebert]

As a stop-gap to at least incorporate single nucleotide variants we could investigate using pyfaidx FastaVariant (https://github.com/mdshw5/pyfaidx):

The FastaVariant class provides a way to integrate single nucleotide variant calls to generate a consensus sequence.