download0 view173
twitter facebook

공공누리This item is licensed Korea Open Government License

Title
Evaluating the effect of database inflation in proteogenomic search on sensitive and reliable peptide identification
Author(s)
이홍란황규백김현우백은옥이상원조윤성
Publication Year
2016-12-22
Abstract
Background: Proteogenomics is a promising approach for various tasks ranging from gene annotation to cancer research. Databases for proteogenomic searches are often constructed by adding peptide sequences inferred from genomic or transcriptomic evidence to reference protein sequences. Such inflation of databases has potential of identifying novel peptides. However, it also raises concerns on sensitive and reliable peptide identification. Spurious peptides included in target databases may result in underestimated false discovery rate (FDR). On the other hand, inflation of decoy databases could decrease the sensitivity of peptide identification due to the increased number of high-scoring random hits. Although several studies have addressed these issues, widely applicable guidelines for sensitive and reliable proteogenomic search have hardly been available.Results: To systematically evaluate the effect of database inflation in proteogenomic searches, we constructed a variety of real and simulated proteogenomic databases for yeast and human tandem mass spectrometry (MS/MS) data, respectively. Against these databases, we tested two popular database search tools with various approaches to search result validation: the target-decoy search strategy (with and without a refined scoring-metric) and a mixture model-based method. The effect of separate filtering of known and novel peptides was also examined. The results from real and simulated proteogenomic searches confirmed that separate filtering increases the sensitivity and reliability in proteogenomic search. However, no one method consistently identified the largest (or the smallest) number of novel peptides from real proteogenomic searches.Conclusions: We propose to use a set of search result validation methods with separate filtering, for sensitive and reliable identification of peptides in proteogenomic search.
Keyword
False discovery rate; Proteogenomic search; Separate false discovery rate analysis; Simulation; Target-decoy approach; Model-based approach
Journal Title
BMC genomics
Citation Volume
17
ISSN
1471-2164
Files in This Item:
There are no files associated with this item.
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
URI
https://repository.kisti.re.kr/handle/10580/14493
http://www.ndsl.kr/ndsl/search/detail/article/articleSearchResultDetail.do?cn=NART78513511
Export
RIS (EndNote)
XLS (Excel)
XML

Browse