In the context of HPLC-MS we worked during the report period in several applied projects on pipelines for quantitative HPLC-MS analysis. In the next years, we will work on more basic research questions, namely on the problem of protein identification resp. protein inference from a set of HPLC-MS measurements. The traditional algorithmic approach for protein identification disregards the inherent information wealth available in a set of mass spectra. By blindly splitting the identification of fragment spectra into several parts, it is not possible that peptide identifcations re-enforce each other and evidence from one spectrum supports the identi cation of a similar one. Further, the MS1 spectra carry more information that is of relevance for identification than just the parent mass of the fragmentation. This information such as further supporting mass positions or retention time information regularly remains either unobserved or is only used out of context and thereby lost for further steps. For more informations about OpenMS click here or read .
People currently working mainly on this topic:
Julianus Pfeuffer: Protein inference with bayesian models