Generating top-down tandem mass spectra (MS/MS) from
complex mixtures
of proteoforms benefits from improvements in fractionation, separation,
fragmentation, and mass analysis. The algorithms to match MS/MS to
sequences have undergone a parallel evolution, with both spectral
alignment and match-counting approaches producing high-quality proteoform-spectrum
matches (PrSMs). This study assesses state-of-the-art algorithms for
top-down identification (ProSight PD, TopPIC, MSPathFinderT, and pTop)
in their yield of PrSMs while controlling false discovery rate. We
evaluated deconvolution engines (ThermoFisher Xtract, Bruker AutoMSn,
Matrix Science Mascot Distiller, TopFD, and FLASHDeconv) in both ThermoFisher
Orbitrap-class and Bruker maXis Q-TOF data (PXD033208) to produce
consistent precursor charges and mass determinations. Finally, we
sought post-translational modifications (PTMs) in proteoforms from
bovine milk (PXD031744) and human ovarian tissue. Contemporary identification
workflows produce excellent PrSM yields, although approximately half
of all identified proteoforms from these four pipelines were specific
to only one workflow. Deconvolution algorithms disagree on precursor
masses and charges, contributing to identification variability. Detection
of PTMs is inconsistent among algorithms. In bovine milk, 18% of PrSMs
produced by pTop and TopMG were singly phosphorylated, but this percentage
fell to 1% for one algorithm. Applying multiple search engines produces
more comprehensive assessments of experiments. Top-down algorithms
would benefit from greater interoperability.