Skip to content

Choose search settings#

This page shows how to choose the digestion rule and search mode for the manifest runner, and how to compare the results. Cleavage specificity and precursor mass window answer different questions. FastaLake exposes them independently in the manifest runner; all combinations use the same database construction and optional molecular additions.

Digestion and search mode#

Option Setting Intended use
--digestion tryptic Two tryptic termini; cleavage after K/R, restricted before P Standard trypsin DDA; default
--digestion semi-tryptic At least one tryptic terminus Inspect peptides with one non-tryptic end
--digestion unspecific No enzyme constraint Peptide processing and endogenous peptide exploration
--search-mode closed Precursor ±20 ppm, fragment ±20 ppm Narrow search; default
--search-mode open Precursor −500 to +100 Da, fragment ±20 ppm Explore mass differences outside declared variable modifications

Protein termini are valid enzymatic boundaries. Search peptide lengths default to 5–50; use --search-min-length and --search-max-length to change them. These are separate from the de novo evidence length filters.

Run a declared comparison#

From the repository root, after installation and preparing the ordinary manifest:

fasta-lake run --manifest samples.tsv --lake lake.fasta \
  --out semitryptic_01 --digestion semi-tryptic --threads 2

fasta-lake run --manifest samples.tsv --lake lake.fasta \
  --out open_01 --search-mode open --threads 2

fasta-lake run --manifest samples.tsv --lake lake.fasta \
  --out unspecific_01 --digestion unspecific \
  --search-min-length 7 --search-max-length 30 --threads 2

Use --validate-only first. Every search writes a complete config.json; the run records selected modes and the Sage binary hash. Open searches retain fixed C carbamidomethylation, disable variable modifications and isotope-error expansion, and disable retention-time prediction. LFQ and generated decoys remain enabled. The mode is a concrete starting configuration, not an automatically optimized instrument or modification model.

These settings target the bundled Sage v0.14.7 release binary, which reports its version as 0.14.6. The installed sage_config function returns a fresh complete JSON dictionary for callers who need to inspect settings. See the Sage configuration documentation and versioned source.

Read the biology carefully#

A non-tryptic end can reflect endogenous processing, sample handling or search error. It does not identify a particular active protease. Report terminal positions, peptide support, acquisition pairing and the database context. An open-search mass difference is not automatically a localized modification.

Compare identification gains and losses on the same acquisitions. Broader settings increase search space, time and often memory; start with a bounded pilot. A narrow-search memory measurement does not establish the cost of a semi-tryptic, unspecific or open search. Independent whole-workflow error calibration remains separate from execution tests.

Executable teaching check#

After the CAMPI example, run:

python examples/searches/run_example.py --campi campi_run_01 \
  --out search_choices_01 --threads 2

The check runs all six combinations on the measured windows with 7–30-aa search peptides. It verifies the target FASTA bytes remain unchanged, acquisition identity is retained, and each route produces study QC. It records outputs and does not assume that a broader setting must yield more identifications.