Hi, I did a webb search and found this:
Proteome Informatics Group: Research activities
The Proteome Informatics Group (PIG) is focusing its activities on the development of software tools and databases for proteomics. Our current topics of research currently cover protein identification and characterization using mass spectrometry, proteomic image analysis, databases, cooperative metaheuristics and their applications to proteomics, development and support in proteome informatics.
New heuristic strategies for automatic protein identification using tandem mass spectrometry data
We are developing a new software, called Popitam, dedicated to protein identification from tandem mass spectrometric data. This new approach exploits a graph structure in association with database searching. The “spectrum graph” allows to store all information coming from the source peak list, in addition to the relationships between the peaks. The graph is then compared with theoretical sequences leading to similarity score for each peptide sequence. As for the graphs currently used for de novo sequencing, the vertices are computed from the spectrum masse/charge ratios and represent masses of potential “ideal” fragments (corresponding to b-ion type charged once). Vertices whose masses differ by the mass value of one or two amino acids are connected by an edge. Sequences can be constructed by moving from one vertex to another one following existing edges. Though, all complete sequences and subsequences that can possibly be built from the experimental MS/MS spectrum are represented in the graph. In Popitam, the aim is to find sections in the graph that explain or suite well theoretical peptides from the database. Mutations and modifications (included unsuspected modifications) can be simulated by combining different sections of the graph in order to maximally cover the theoretical peptide.
It appears to be a method of compiling and comparing statistics.