pefftacular¶

pefftacular reads and writes PEFF (PSI Extended FASTA Format) files in pure Python, with no dependencies.

What is PEFF?¶

PEFF is the HUPO Proteomics Standards Initiative's extension of FASTA. A PEFF file is still a FASTA file: every entry is a > description line followed by the sequence. PEFF adds two things:

a file header of # Key=value lines that describes each source database (name, version, prefix, entry count, custom keys), and
structured \Key=value pairs on each description line, so an entry can carry gene and protein names, taxonomy, sequence variants, post-translational modifications, signal peptides, disulfide bonds and whole proteoforms.

# PEFF 1.0
# //
# DbName=SwissProt
# Prefix=sp
# DbVersion=2026_03
# DbSource=https://www.uniprot.org
# NumberOfEntries=1
# SequenceType=AA
# //
>sp:P12345 \PName=Example protein \GName=EXMP \VariantSimple=(5|C|dbSNP) \ModResUnimod=(13|UNIMOD:21|Phospho)
MKTAYIAKQRQISFVKSHFSRQ

Search engines that support PEFF, such as Comet, can read these annotations and search variant and modified peptides directly from the database.

Why pefftacular?¶

Typed data, not strings. Headers, entries and every annotation type are frozen dataclasses. You get entry.mod_res_unimod[0].positions, not a regex over the header.
Streaming or eager. read_peff loads a whole file; PeffReader yields one entry at a time.
Permissive reading, strict when you want it. Spec violations are reported as PeffWarning and the data is still returned. Structural failures raise PeffParseError with the line number, the offending text and a hint.
Round-trip writing with the backslash escaping PEFF requires.
Zero dependencies, Python 3.12+.

Install¶

pip install pefftacular
# or, in a uv project
uv add pefftacular

30-second example¶

Build an entry with a variant and a phosphorylation site, write it, and read it back:

from pefftacular import (
    DatabaseHeader,
    FileHeader,
    ModResUnimod,
    SequenceEntry,
    VariantSimple,
    read_peff,
    write_peff,
)

header = FileHeader(
    peff_version="1.0",
    databases=(
        DatabaseHeader(
            prefix="sp",
            db_name="SwissProt",
            db_version="2026_03",
            db_sources=("https://www.uniprot.org",),
            number_of_entries=1,
            sequence_type="AA",
        ),
    ),
)

entry = SequenceEntry(
    prefix="sp",
    db_unique_id="P12345",
    sequence="MKTAYIAKQRQISFVKSHFSRQ",
    pname="Example protein",
    gname="EXMP",
    variant_simple=(VariantSimple(position=5, new_amino_acid="C", tag="dbSNP"),),
    mod_res_unimod=(ModResUnimod(positions=(13,), accession="UNIMOD:21", name="Phospho"),),
)

write_peff(header, [entry], "example.peff")

header, entries = read_peff("example.peff")
for e in entries:
    print(e.db_unique_id, e.gname, len(e.sequence))
    for mod in e.mod_res_unimod:
        print(mod.name, "at", mod.positions)
# P12345 EXMP 22
# Phospho at (13,)

print(entries[0] == entry)
# True

Where next¶

Getting started: reading, writing, the header and entry model, errors.
Annotations: variants, modifications, processing, disulfide bonds, proteoforms and custom keys.
From FASTA and UniProt: turn a plain FASTA file into PEFF.
API reference: every public class and function.
Using an AI coding assistant? Point it at llms.txt for a compact API guide.

fastatacular reads and writes plain FASTA with the same API shape.
peptacular handles peptide sequences, ProForma notation, digestion and mass calculation.
unimodpy and psimodpy look up the UNIMOD and PSI-MOD accessions used in ModResUnimod and ModResPsi.