Dataset Information

High quality MS/MS spectrum prediction for data-dependent and -independent acquisition data analysis

ABSTRACT: Peptide fragmentation spectra are routinely predicted in the interpretation of mass spectrometry-based proteomics data. Unfortunately, the generation of fragment ions is not well enough understood to estimate fragment ion intensities accurately. Here, we demonstrate that machine learning can predict peptide fragmentation patterns in mass spectrometers with accuracy within the uncertainty of the measurements. Moreover, analysis of our models reveals that peptide fragmentation depends on long-range interactions within a peptide sequence. We illustrate the utility of our models by applying them to the analysis of both data-dependent and data-independent acquisition datasets. In the former case, we observe a significant increase in the total number of peptide identifications at fixed false discovery rate. In the latter case we demonstrate that the use of predicted MS/MS spectra is equivalent to the use of spectra from experimentallibraries, indicating that fragmentation libraries for proteomics are becoming obsolete.

INSTRUMENT(S):

ORGANISM(S): Homo Sapiens (human)

TISSUE(S): Blood Plasma

SUBMITTER: Shivani Tiwary

LAB HEAD: Juergen Cox; Peter Cimermancic

PROVIDER: PXD010382 | Pride | 2019-03-13

REPOSITORIES: Pride

ACCESS DATA

Dataset's files

Source:

			Action	DRS
	20161213_NGHF_DBJ_SA_Exp3A_HeLa_1ug_60min_15000_01.raw	Raw
	20161213_NGHF_DBJ_SA_Exp3A_HeLa_1ug_60min_15000_02.raw	Raw
	20161213_NGHF_DBJ_SA_Exp3A_HeLa_1ug_60min_15000_03.raw	Raw
	20171102_091300_Test-InSilico-20171030-real_Report.zip	Other
	20171102_091314_Test-Insilico-20171030-2d_Report.zip	Other

Items per page:

1 - 5 of 34

Publications

High-quality MS/MS spectrum prediction for data-dependent and data-independent acquisition data analysis.

Tiwary Shivani S Levy Roie R Gutenbrunner Petra P Salinas Soto Favio F Palaniappan Krishnan K KK Deming Laura L Berndl Marc M Brant Arthur A Cimermancic Peter P Cox Jürgen J

Nature methods 20190527 6

Peptide fragmentation spectra are routinely predicted in the interpretation of mass-spectrometry-based proteomics data. However, the generation of fragment ions has not been understood well enough for scientists to estimate fragment ion intensities accurately. Here, we demonstrate that machine learning can predict peptide fragmentation patterns in mass spectrometers with accuracy within the uncertainty of measurement. Moreover, analysis of our models reveals that peptide fragmentation depends on ...[more]

PMID: 31133761

Similar Datasets

Project description:Background Plasma proteins can be modified by addition of sugar residues by non-enzymatic glycation as well as glycosylation. While glycation is a hallmark of hyperglycemia, glycosylation is a complex, enzyme-catalyzed post-translational modification by heterogeneous sugar chains and is present on a majority of plasma proteins. N-linked glycosylation occurs on asparagine residues occurring predominantly on a canonical N-glycosylation motif (Asn-X-Ser/Thr) although non-canonical N-glycosylation motifs Asn-X-Cys/Val have also been reported to be glycosylated less frequently. Albumin is the most abundant protein in human plasma whose glycation has been implicated in end-organ damage in advanced diabetes mellitus, and as such, has both diagnostic and therapeutic implications. However, albumin has long been regarded as a non-glycosylated protein owing to the absence of canonical motifs. Albumin contains two non-canonical N-glycosylation motifs, of which one was recently reported to be glycosylated with two possible attached glycans. Methods To investigate N-linked glycosylation of abundant serum proteins in greater detail, we passed healthy donor samples through a multiple affinity removal spin (MARS14) column followed by trypsin digestion of bound proteins and subsequent glycopeptide enrichment by size-exclusion chromatography (SEC) or mixed-mode anion-exchange (MAX), in parallel. Global analysis of intact glycopeptides was performed by LC-MS/MS to evaluate potential glycosylation at canonical as well as non-canonical sites. PNGase F treatment was carried out to confirm N-linked glycosylation at non-canonical sites Asn-X-Cys/Val. Glycopeptides on albumin at non-canonical sites were further characterized by MS3 fragmentation from immunoprecipitated albumin to confirm the attached glycans. To assess if the observed glycosylation was a general phenomenon, targeted analysis using parallel reaction monitoring (PRM) was carried out on an independent set of twenty human serum samples. Finally, to test whether albumin glycosylation is unique to human albumin or is a general phenomenon, bovine and rabbit serum albumins were subjected to MAX chromatography and LC-MS/MS analysis. Results We report that human albumin is modified by N-linked glycosylation at two sites, Asn68(-Glu-Val) and Asn123-(Glu-Cys), both of which occur in non-canonical N-glycosylation motifs. We detected N-glycopeptides at both sites bearing four complex mono- and bi-antennary N-linked glycans, with Hex5HexNAc4NeuAc2 and Hex5HexNAc4NeuAc1 being the most abundant glycans. These findings were validated by analyzing deglycosylated peptides and MS3-based fragmentation of glycopeptides. Glycosylation at both sites was confirmed by targeted MS in twenty donor samples. Finally, we report that the Asn residue on bovine and rabbit serum albumins that corresponds to the highly conserved Asn123 in human albumin is also N-glycosylated with complex sialylated glycans similar to those observed in human albumin. Conclusions Albumin is a conserved glycoprotein with multiple N-linked glycosylation sites. This glycosylation of albumin has potential functional implications and applications in medicine.