False Discovery Rate Estimation for Hybrid Mass Spectral Library Search Identifications in Bottom-up Proteomics

Published

July 31, 2019

Author(s)

Meghan C. Burke, Zheng Zhang, Yuri A. Mirokhin, Dmitrii V. Tchekhovskoi, Yuxue Liang, Stephen E. Stein

Abstract

We present a method for FDR estimation of mass spectral library search identifications made by a recently developed method for peptide identification, the hybrid search, based on an extension of the target-decoy approach. In addition to estimating confidence for a given identification, this allows users to compare and integrate identifications from the hybrid mass spectral library search method with other peptide identification methods, such as a sequence database-based method. In addition to a score, each hybrid score is associated with a ‘ΔMass’ value, which is the difference in mass of the search and library peptide, which can correspond to the mass of a modification. We explored the relation between FDR and ΔMass using 100 concatenated random decoy libraries, and discovered that a small number of ΔMass values were especially likely to result from decoy searches. Using these values, FDR values could be adjusted for these specific values, and a reliable FDR generated for any ΔMass value. Finally, using this method, we find and examine common, reliable identifications made by the hybrid search for a range of proteomic studies.

Citation

ACS Journal of Proteome Research

Pub Type

Journals

Download Paper

https://doi.org/10.1021/acs.jproteome.8b00863

Local Download

Keywords

Peptide mass spectral library, hybrid mass spectral library search, false discovery rate, target-decoy approach

Citation

Burke, M. , Zhang, Z. , Mirokhin, Y. , Tchekhovskoi, D. , Liang, Y. and Stein, S. (2019), False Discovery Rate Estimation for Hybrid Mass Spectral Library Search Identifications in Bottom-up Proteomics, ACS Journal of Proteome Research, [online], https://doi.org/10.1021/acs.jproteome.8b00863 (Accessed September 13, 2024)