Implementations of knapsack algorithms for metabolite identification

Metabolite identification is an important problem in Bioinformatics. Given the mass of a molecule it needs to identify the number of various atoms that can compose it. This type of problem relates to the larger type of problems known as the Knapsack Problem. Therefore, three algorithms that can solve it were implemented: Exhaustive Search, Dynamic Programming Bellman’s algorithm[3] and the algorithms used for the SIRIUS system[4] which have been introduced in [5]. The implementations have been evaluated against each other on in silico data and the results reflect that the SIRIUS-like knapsack algorithm produces better output at much faster speeds than more classical algorithms. The system has several other functions such as generating test data from molecule formulas, running against a dataset read from file and filtering results. It can also be run from the console with intuitive commands that walk the user through entering the data and parameters needed.