Wang Jianqi, Zhang Yajie, Yu Yonghao
Department of Biochemistry, The University of Texas Southwestern Medical Center, Dallas, TX, 75390, USA,
J Am Soc Mass Spectrom. 2015 Jul;26(7):1077-84. doi: 10.1007/s13361-015-1120-3. Epub 2015 Apr 21.
A search engine that discovers more peptides reliably is essential to the progress of the computational proteomics. We propose two new scoring functions (L- and P-scores), which aim to capture similar characteristics of a peptide-spectrum match (PSM) as Sequest and Comet do. Crescendo, introduced here, is a software program that implements these two scores for peptide identification. We applied Crescendo to test datasets and compared its performance with widely used search engines, including Mascot, Sequest, and Comet. The results indicate that Crescendo identifies a similar or larger number of peptides at various predefined false discovery rates (FDR). Importantly, it also provides a better separation between the true and decoy PSMs, warranting the future development of a companion post-processing filtering algorithm.
一个能够更可靠地发现更多肽段的搜索引擎对于计算蛋白质组学的进展至关重要。我们提出了两种新的评分函数(L评分和P评分),旨在像Sequest和Comet那样捕捉肽段-谱匹配(PSM)的相似特征。这里介绍的Crescendo是一个为肽段鉴定实现这两种评分的软件程序。我们将Crescendo应用于测试数据集,并将其性能与广泛使用的搜索引擎(包括Mascot、Sequest和Comet)进行比较。结果表明,在各种预定义的错误发现率(FDR)下,Crescendo鉴定出的肽段数量相似或更多。重要的是,它还能在真实和诱饵PSM之间提供更好的区分,这为后续开发配套的后处理过滤算法提供了依据。