Evaluation of database search programs for accurate detection of neuropeptides in tandem mass spectrometry experiments

Programs to identify proteins in tandem mass spectrometry experiments are not optimized to identify neuropeptides and other peptides resulting from the processing of prohormones. This is due to the unique characteristics of neuropeptides including release after complex processing of prohormones, potentially intense post-translational modifications and their small size. The aims of this study were: (1) to evaluate the strengths and limitations of different tandem mass spectra search algorithms to detect neuropeptides and other peptides resulting from prohormone processing; (2) to evaluate the impact of mass spectrometry factors such as charge on the identification of these peptides; and (3) to offer guidelines to obtain the most comprehensive and accurate survey of the prohormone peptides of a sample. Three software database search programs, OMSSA, X!Tandem and Crux, were applied to identify neuropeptides from in silico produced mass spectra. The spectra were simulated from a database of 7850 mouse peptides from 92 prohormones. For each peptide, spectra were simulated with either +1, +2 and +3 precursor charge states, and +1 charged b and y product ions including single water and/or ammonia loss depending on amino acid composition. The spectra were searched against the mouse database and a rat database including 7647 neuropeptides. OMSSA, X!Tandem and Crux correctly detected 98.9%, 93.9% and 88.7% of the peptides, respectively, at the comparable significance E- or p-value < 1 x 10-6. Scoring only b- or y-ion series significantly reduced peptide identification for both OMSSA and X!Tandem. At E-value < 1 x 10-6, 50.8% and 55.3% of peptides were correctly identified by both algorithms using b- and y-ion series, respectively. Furthermore, availability of only b-ion, y-ion series and 50% random ions for peptide identification had in general minor influence on the scoring functions of OMSSA and Crux. The comparatively weaker performance of X!Tandem suggests that the corresponding scoring function favors continuity of ions. The charge state had minor effect on the detection of neuropeptides. Unlike Crux and X!Tandem, OMSSA was negatively influenced by the presence of additional peaks in the spectra at higher precursor charge states. The sensitivity of either program to detect small neuropeptides (< 10 amino acids in length) was limited. This is particularly troublesome given the large number of neuropeptides that are small. Peptide identification by X!Tandem across species suggests that the position of the mismatch in the sequence is critical when using non-specific species databases. These results indicate that alternative algorithmic specifications and implementations must be developed to optimize the detection of neuropeptides.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Bioinformatics

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2012

Author dc:creator

• Akhtar, Malik Nadeem

Contributors dc:contributor

• Rodriguez-Zas, Sandra L.

• database search program
• prohormones
• Crux
• X!Tandem
• Open Mass Spectrometry Search algorithm (OMSSA)

Statement dc:rights

• Copyright 2012 Malik Akhtar

Language dc:language

eng

Handle dc:identifier

http://hdl.handle.net/2142/34590

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/34590