comet-search-run

name	comet-search-run
description	Run Comet batch or real-time searches, interpret output files, and understand search parameters. Use when executing searches, reading results, or configuring search settings.

Comet Search Run

Batch search

Comet.exe -D<database.idx> <input.raw>
# or with a FASTA database:
Comet.exe -D<database.fasta> <input.raw>

-D specifies the database. The .idx suffix triggers fragment ion index mode (FI_DB); .fasta triggers standard FASTA search mode.

Additional flags:

• -P<comet.params> — use a specific params file (default: comet.params in working dir)
• -N<output_basename> — override output file base name

Create .idx file for Real-time search (RTS)

Comet.exe -D<database.fasta> -i
Comet.exe -D<database.fasta> -k

• the "-i" command line option creates a fragment ion index plain peptide database with .idx extension
• the "-j" command line option creates a peptide index database with .idx extension

Real-time search (RTS)

RealtimeSearch.exe <query.raw> <ms1_reference.raw> <database.idx> <num_threads>

• query.raw — file containing MS2 spectra to search
• ms1_reference.raw — file containing MS1 scans for RT alignment (can be the same file)
• database.idx — fragment ion index database
• num_threads — number of parallel search threads (typically 20)

Output is written to rts.out in the working directory.

Output file format (rts.out)

Each identified MS2 scan:

MS2 <scan>  <prevAA>.<peptide>.<nextAA>  <xcorr>  <evalue>  z <charge>  exp <exp_mass>  calc <calc_mass>  AScore <ascore>  Sites '<sites>'  <ms> ms  prot '<protein>'

Example:

MS2 1325  K.HAVSEGTK.A  1.8840  1.68E-08  z 2  exp 827.4132  calc 827.4137  AScore 0.00  Sites ''  8 ms  prot 'sp|O60814|H2B1K_HUMAN Histone H2B...'

Modification notation: [mass_shift] inline, e.g. S[79.9663] = phosphoserine, M[15.9949] = oxidized Met. N-terminal mod: n[mass], C-terminal mod: c[mass].

Output also includes:

• Search histogram (ms per spectrum distribution)
• <= 5ms / > 5ms / > 10ms breakdown
• 5 slowest MS2 runs
• initialize search elapsed time — index load time
• search elapsed time — parallel search time with avg ms/spec and Hz

Database types

Suffix	Type constant	Description
.idx	FI_DB	Fragment ion index (pre-built, fast RTS)
.fasta	FASTA_DB	Standard FASTA, searched directly
.idx (peptide)	PI_DB	Peptide index variant

Key search parameters (comet.params)

• peptide_mass_tolerance / peptide_mass_units — precursor mass window
• fragment_bin_tol / fragment_bin_offset — fragment ion binning
• enzyme_number — 0 = no enzyme, 1 = trypsin, etc.
• allowed_missed_cleavage
• peptide_length_range — e.g. 8 25 for MHC peptides
• num_threads — overridden by RTS command-line argument
• num_results — hits to report per spectrum
• minimum_xcorr — score cutoff

g_pvProteinNameCache (indexed DB)

At index load time, all protein names from the .idx file are cached in g_pvProteinNameCache (an unordered_map<comet_fileoffset_t, string>). For human.target-decoy.fasta.idx this is ~40,858 entries, ~7 MB RAM. RTS protein lookups are then in-memory hashtable hits with no per-spectrum file I/O.