IEDB/PEPMatch - Githubissues

Author: Daniel Marrama

Peptide search against a reference proteome, or sets of proteins, with residue subtitutions.

Two step process: preprocessing and matching.

Preprocessed data is stored in a SQLite or pickle format and only has to be performed once.

As a competition to improve tool performance, we created a benchmarking framework with instructions here.

Requirements

Python 3.7+
Pandas
NumPy
Biopython

Installation

pip install pepmatch

Inputs

Preprocessor

proteome - Path to proteome file to search against.\ k - k-mer size to break up proteome into.\ preprocessed_format - SQLite ("sqlite") or "pickle".\ preprocessed_files_path - (optional) Directory where you want preprocessed files to go. Default is current directory.\ gene_priority_proteome - (optional) Subset of proteome with prioritized protein IDs.\

Matcher

query - Query of peptides to search either in .fasta file or as a Python list.\ proteome_file - Name of preprocessed proteome to search against.\ max_mismatches - Maximum number of mismatches (substitutions) for query.\ k - (optional) k-mer size of the preprocessed proteome. If no k is selected, then a best k will be calculated and the proteome will be preprocessed\ preprocessed_files_path - (optional) Directory where preprocessed files are. Default is current directory.\ best_match - (optional) Returns only one match per query peptide. It will output the best match.\ output_format - (optional) Outputs results into a file (CSV, XLSX, JSON, HTML) or just as a dataframe.\ output_name - (optional) Specify name of file for output. Leaving blank will generate a name.

Note: For now, due to performance, SQLite is used for exact matching and pickle is used for mismatching.

Note: PEPMatch can also search for discontinuous epitopes in the residue:index format. Example:

"R377, Q408, Q432, H433, F436, V441, S442, S464, K467, K489, I491, S492, N497"

Command Line Example

# exact matching example
pepmatch-preprocess -p human.fasta -k 5 -f sql
pepmatch-match -q peptides.fasta -p human.fasta -m 0 -k 5

# mismatching example
pepmatch-preprocess -p human.fasta -k 3 -f pickle
pepmatch-match -q neoepitopes.fasta -p human.fasta -m 3 -k 3

Exact Matching Example

from pepmatch import Preprocessor, Matcher

Preprocessor('proteomes/human.fasta').sql_proteome(k = 5) 

Matcher( # 0 mismatches, k = 5
  'queries/mhc-ligands-test.fasta', 'proteomes/human.fasta', 0, 5
).match()

Mismatching Example

from pepmatch import Preprocessor, Matcher

Preprocessor('proteomes/human.fasta').pickle_proteome(k = 3)

Matcher( # 3 mismatches, k = 3
  'queries/neoepitopes-test.fasta', 'proteomes/human.fasta', 3, 3
).match()

Parallel Matcher Example

To run a job on multiple cores, use the ParallelMatcher class. The n_jobs parameter specifies the number of cores to use.

from pepmatch import Preprocessor, ParallelMatcher 

Preprocessor('proteomes/betacoronaviruses.fasta').pickle_proteome(k = 3)

ParallelMatcher(
  query='queries/coronavirus-test.fasta',
  proteome_file='proteomes/betacoronaviruses.fasta',
  max_mismatches=3,
  k=3,
  n_jobs=2
).match()

Best Match Example

from pepmatch import Matcher
Matcher(
  'queries/milk-peptides-test.fasta', 'proteomes/human.fasta', best_match=True
).match()

The best match parameter without k or mismatch inputs will produce the best match for each peptide in the query, meaning the match with the least number of mismatches, the best protein existence level, and if the match exists in the gene priority proteome. No preprocessing beforehand is required, as the Matcher class will do this for you to find the best match.

Outputs

As mentioned above, outputs can be specified with the output_format parameter in the Matcher class. The following formats are allowed: dataframe, tsv, csv, xlsx, json, and html.

If specifying dataframe, the match() method will return a pandas dataframe which can be stored as a variable:

df = Matcher(
  'queries/neoepitopes-test.fasta', 'proteomes/human.fasta', 3, 3, output_format='dataframe'
).match()

Citation

If you use PEPMatch in your research, please cite the following paper:

Marrama D, Chronister WD, Westernberg L, et al. PEPMatch: a tool to identify short peptide sequence matches in large sets of proteins. BMC Bioinformatics. 2023;24(1):485. Published 2023 Dec 18. doi:10.1186/s12859-023-05606-4

IEDB / PEPMatch

readme