RAId{\_}DbS: Method for Peptide ID using Database Search with Accurate Statistics

The key to proteomics studies, essential in systems biology, is peptide identification. Under tandem mass spectrometry, each spectrum generated consists of a list of mass/charge peaks along with their intensities. Software analysis is then required to identify from the spectrum peptide candidates that best interpret the spectrum. The library search, which compares the spectral peaks against theoretical peaks generated by each peptide in a library, is among the most popular methods. This method, although robust, lacks good quantitative statistical underpinning. As we show, many library search algorithms suffer from statistical instability. The need for a better statistical basis prompted us to develop RAId{\_}DbS. Taking into account the skewness in the peak intensity distribution while scoring peptides, RAId{\_}DbS provides an accurate statistical significance assignment to each peptide candidate. RAId{\_}DbS will be a valuable tool especially when one intends to identify proteins through peptide identifications.