Figure 2.

Data analysis workflow and achieved data reduction in representative datasets. (a) Outline of data analysis pipeline. Raw data from LC-ESI-MS/MS experiments is submitted to the data analysis pipeline. In preparation for subsequent steps, raw data is converted into an open mass spectrometry format (.mzML) and centroided; retention time alignment between the UV sample and the nonirradiated control is performed. Next, the overall amount of data is reduced by removing MS/MS spectra of confidently identified noncross-linked peptides (ID filter) and species appearing in both UV-irradiated sample and control with comparable intensities (XIC filter with RNP^xlXIC). Lastly, the RNP^xl tool removes MS/MS spectra with small precursors masses (low m/z filter) and residual short oligonucleotides (fractional mass filter). For the key steps in data analysis, RNP^xl creates precursor mass variants, submits data into the search engine, and summarizes the search results. (b) Results of data analysis procedure for a single dataset of yeast RNA-binding proteins. XIC, ID, and fractional mass filter excluded approximately two thirds (67%) of the overall 9,728 fragment spectra. Additionally, 16% of the spectra could be disregarded as these did not yield any database search result for a cross-linked heteroconjugate. Of the remaining 17% potential cross link candidates, 14% had a low score (E-value ≥ 0.01) yielding a final list of 317 (3%) potential cross link candidates for manual validation.