Dipartimento di Scienze Agrarie, degli Alimenti e dell'Ambiente and CSRA, Centro Servizi di Ricerca Applicata, Università degli Studi di Foggia, Via Napoli, 25, 71122 Foggia, Italy.
Dipartimento di Scienze Agrarie, degli Alimenti e dell'Ambiente and CSRA, Centro Servizi di Ricerca Applicata, Università degli Studi di Foggia, Via Napoli, 25, 71122 Foggia, Italy.
Talanta. 2017 Mar 1;164:684-692. doi: 10.1016/j.talanta.2016.10.102. Epub 2016 Nov 1.
A fundamental issue in proteomics is the peptide identification by database searching and the assessment of the goodness of fit between experimental and theoretical data. Despite the different number of ways to measure the quality of search results, the definition of a scoring criterion is still highly desirable in ion-trap based proteomics. Indeed, in order to fully take advantage of a low resolution MS/MS dataset, it is essential to strike a balance between greater information capture and reduced number of incorrect peptide assignments. In addition, the development of user-specified rules is a crucial aspect when very similar proteins of the same family are analyzed in order to infer the origin species. In this study, a post-processing validation scheme is provided for the evaluation of proteomic data in shot-gun ion-trap proteomics, when a flexible database searching based on the error tolerant mode is adopted in combination with a low-specificity enzyme to maximize sequence coverage. To validate peptide assignments, we used standard β-casein digested with trypsin/chymotrypsin or trypsin alone and the popular search engine MASCOT to identify the relevant (known) peptide sequences. A linear combination between peptide ion score and normalized delta score (i.e. the difference between the best and the second best ion score, divided by the best score) is proposed to increase the accuracy in sequence assignments from low-resolution tandem mass spectra. Finally, the optimized post-processing database validation was successfully applied to the direct analysis of milk tryptic/chymotryptic digests of different origin, without resorting to two-dimensional electrophoresis that is usually performed for protein separation in ion-trap proteomics. The identification of species-specific amino acidic sequences among the validated peptide spectrum matches has allowed to fully discriminate between the animal species, so evaluating accurately the milk authenticity.
蛋白质组学中的一个基本问题是通过数据库搜索对肽进行鉴定，并评估实验数据和理论数据之间的拟合程度。尽管有不同的方法来衡量搜索结果的质量，但在基于离子阱的蛋白质组学中，仍然非常需要定义评分标准。事实上，为了充分利用低分辨率 MS/MS 数据集，在增加信息量和减少错误肽分配之间取得平衡至关重要。此外，当分析同一家族非常相似的蛋白质以推断起源物种时，制定用户指定的规则是一个关键方面。在这项研究中，提供了一种后处理验证方案，用于评估 shotgun 离子阱蛋白质组学中的蛋白质组数据，当采用基于容错模式的灵活数据库搜索并结合低特异性酶以最大化序列覆盖度时。为了验证肽分配，我们使用了标准的β-酪蛋白，用胰蛋白酶/糜蛋白酶或胰蛋白酶单独消化，并使用流行的搜索引擎 Mascot 来识别相关（已知）肽序列。提出了肽离子得分和归一化差值得分（即最佳和第二最佳离子得分之间的差异，除以最佳得分）之间的线性组合，以提高从低分辨率串联质谱中获得序列分配的准确性。最后，优化后的后处理数据库验证成功应用于不同来源的牛奶胰蛋白酶/糜蛋白酶消化物的直接分析，而无需进行二维电泳，二维电泳通常用于离子阱蛋白质组学中的蛋白质分离。在经过验证的肽谱匹配中鉴定出物种特异性氨基酸序列，从而能够完全区分动物物种，从而准确评估牛奶的真实性。
