FASTM API (EBI)

FASTM API (EBI) performs sequence similarity searches using peptide- and oligonucleotide-derived sequences to identify homologous proteins and nucleotides for proteomic, genomic, and evolutionary analyses.

Key Features:

• FASTF Algorithm: Compares ordered mixtures of peptides (typically from Edman degradation after CNBr cleavage) against protein databases using greedy deconvolution and alignment-probability optimality with an empirical correction to probability estimates, with statistical estimates accurate within a factor of 1000 and requiring ~25% more sequence data than FASTS for equivalent sensitivity.
• FASTM Algorithm: Compares ordered sets of peptides to protein sequence databases and ordered oligonucleotides to nucleotide sequence databases to support both proteomic and genomic queries.
• FASTS Algorithm: Searches unordered peptide sets (typically from mass spectrometry) by evaluating all peptide orderings, employing a heuristic FASTA comparison strategy and alignment-probability optimality with an empirical correction, achieving statistical accuracy within a factor of 10 and detecting homologues with ≥50% identity using 15–20 total residues across three or four peptides.
• Database support: Performs searches against protein and nucleotide sequence databases.
• Statistical scoring: Uses alignment probability as the criterion for optimal alignments with empirical corrections to theoretical probabilities for statistical estimates.

Scientific Applications:

• Proteomic identification in unsequenced organisms: Identifies homologous proteins in taxa lacking genome sequences, including homologues that diverged 100–500 million years ago.
• Mass spectrometry-based peptide analysis: Enables identification of proteins from unordered peptide fragments derived by mass spectrometry using FASTS.
• Edman-degradation-based sequencing: Enables identification of proteins from ordered peptide mixtures produced by Edman degradation after CNBr cleavage using FASTF.
• Evolutionary and comparative genomics: Supports studies of ancient protein functions and sequence divergence across deep evolutionary timescales (100–500 million years).

Methodology:

Computational methods include greedy-heuristic deconvolution for ordered mixed peptides (FASTF), exhaustive evaluation of peptide orderings for unordered peptides (FASTS), a heuristic FASTA comparison strategy to accelerate searches, use of alignment probability as the optimality criterion, and empirical correction of theoretical probability estimates; FASTM performs direct comparisons of ordered peptides to protein databases and ordered oligonucleotides to nucleotide databases.