Names used in FastaLake

A biological sample, a database entry and a reported protein group describe different things. These definitions follow the workflow from input to result.

Term	Meaning	Where you see it
Specimen	Biological material linking proteomic and matched DNA/RNA measurements. One specimen can have several acquisitions.	Molecular source manifest
Acquisition	One mass-spectrometry measurement, supplied as one mzML file with its predictions. The manifest's sample column names acquisitions.	samples.tsv
Protein reservoir	The prepared collection of complete protein sequences from which candidates are retrieved. Also called the reference lake in commands.	--lake
De novo query	An eligible predicted peptide retained for exact lookup. It is selection evidence, before database-search acceptance.	Prediction CSV
Study-supported sequence space	Reservoir proteins containing a retained query from the declared study. Also called the evidence lake in output files.	evidence.fasta
Acquisition-specific candidate set	Proteins retrieved from the study-supported space by one acquisition's retained queries.	sample_lake.fasta
Selection parsimony	Deterministic razor assignment of mapped predicted peptides to candidates; proteins receiving an assignment are retained in full.	Rust parsimony_engine
Selected acquisition-specific database	Complete proteins retained after selection parsimony. This is the de novo baseline.	<sample>_razor.fasta
Final target FASTA	Proteins supplied to the search, including any requested molecular additions. Sage generates decoys separately.	Search config.json
Accepted peptide	A target peptide passing the declared search-confidence threshold. Canonical sequence summaries remove modification notation and treat I/L as equivalent.	Peptide evidence tables
Study-level protein group	A reporting representative and the accepted peptides assigned to it across acquisitions. Shared evidence can leave biological origin ambiguous.	study_groups.tsv
Protein-group quantity	The sum or directLFQ estimate for a study group in an acquisition. An unquantified cell remains missing.	study_group_matrix.tsv

Two kinds of sequence identity

Peptide lookup treats isoleucine and leucine as equivalent. The prepared lake's protein hashes preserve the original I/L distinction: one matching peptide can occur in several distinct protein sequences. See identity rules.

Selection and reporting

Before searching, selection parsimony uses predicted peptides to choose a compact FASTA. After searching, study grouping uses accepted evidence to build a common reporting dictionary. They use different evidence and tie rules.

The workflow shows both operations. The function map points to their implementations.

Where molecular additions enter

Appending specimen-matched DNA/RNA-supported sequences after selection differs from pooling assembly proteins into the reservoir before lookup. Report where each source entered; “uses metagenomics” alone does not specify the method.