proteomics-identification

proteomics-identification

When to use

Confidence filtering uses qvalue, q-value, q_value, PEP, pep or fdr in that order. The default threshold is 0.01. Use existing search-engine tables; this skill does not search raw spectra.

Use from a step

from skills._sdk.notebook import load_skill, write_output
library = load_skill('proteomics-identification')
data = library.demo_data(random_state=42)
result = library.filter_identifications(data, n_spectra=1000)
write_output(result, 'tables/peptides.csv')

For real data, use read_input and pass any read_table helper as reader=. The executable examples/example_step.py also checks the result and writes a Figure.

API

read_table(path: str | Path) -> pd.DataFrame

Read peptide CSV/TSV; pass this function as reader= to read_input.

:param path: Peptide table; txt and tsv suffixes select tab separation. :returns: Table with common MaxQuant column names normalized. :raises OSError: The file cannot be read.

filter_identifications(data: pd.DataFrame, *, fdr_threshold: float=0.01, n_spectra: int | None=None) -> pd.DataFrame

Filter peptide confidence values and return a new table.

:param data: Existing peptide/protein rows with optional qvalue, q-value, q_value, PEP, pep or fdr. :param fdr_threshold: CLI default 0.01; PEP thresholding is not a global FDR estimate. :param n_spectra: Total spectra; CLI default None uses retained PSM count, not an observed identification rate. :returns: Filtered rows with the actual confidence column and summary in attrs. :raises ValueError: Threshold or spectrum count is invalid.

run_info(table: pd.DataFrame, *, keep: bool=True) -> dict

Read identification diagnostics.

:param table: Filtered peptide table. :param keep: True preserves attrs; False removes diagnostics. :returns: A separate dictionary with filter provenance and summary. :raises TypeError: The input is not a DataFrame.