Estimating the computational limits of detection of microbial non-model organisms

Mass spectrometry has become a key instrument for proteomic studies of single bacteria as well as microbial communities. However, the identification of spectra from MS/MS experiments is still challenging, in particular for non-model organisms. Due to the limited amount of related protein reference sequences, underexplored organisms often remain completely unidentified or their spectra match to peptides of uncertain degree of relation. Alternative strategies such as error-tolerant spectra searches or proteogenomic approaches may reduce the amount of unidentified spectra and lead to peptide matches on more related taxonomic levels. However, to what extent these strategies may be successful is difficult to judge prior to an MS/MS experiment. In this contribution, we introduce a method to estimate the suitability of databases of interest. Further, it allows estimating the possible influence of error-tolerant searches and proteogenomic approaches on databases of interest with respect to the number of unidentified spectra and the taxonomic distances of identified spectra. Furthermore, we provide an implementation of our approach that supports experimental design by evaluating the benefit and need of different search strategies with respect to present databases and organisms under study. We provide several examples which highlight the different effects of additional search strategies on databases and organisms with varying amount of known relative species available.