Bioinformatics tools for proteomics research

蛋白质组学涉及复杂生物系统中所有蛋白质的特征，可以提供对疾病的重要洞察，识别用于诊断和个性化药物的生物标记物。今天，蛋白质的表征是通过质谱学实验进行的，然后用软件进行数据分析。然而，数据分析仍然是这一进程中的一个主要瓶颈。为了应对这一挑战，我开发了创新的算法和分析软件，世界各地来自学术界和工业界的研究实验室的数千名研究人员经常使用这些算法和分析软件。 这项拟议研究的长期目标是为生物和健康研究人员提供一个强大的软件平台，用于充分利用质谱学来表征蛋白质。在新算法的支持下，该平台的精确度将大大提高，并将具有内置的质量控制，以保证高质量的结果。这样的平台将使许多生物和健康应用中的发现过程得以实现并加速。在这项研究提案中，将追求四个短期目标和分任务： (1)更准确的从头测序：从头测序识别任何数据库中不存在的新肽。我们的研究将解决目前方法的准确性不足和质量控制不足的问题。 (2)低误发现率(FDR)的新数据库搜索方法：对于数据库中存在的多肽，我们将研究新的方法，在减少FDR的同时识别更多的多肽。 (3)翻译后修饰(PTM)：翻译后修饰及其数量变化是非常有前景的生物标志物。我们将开发有效的计算方法来识别和量化修饰的多肽。 (4)人类蛋白质组图谱：为了更好地表征复发的人类多肽，我们打算通过分析所有公开的人类质谱学数据来建立人类蛋白质组图谱。 要实现这些目标，需要创新的算法。本提案中所述的研究活动将使有效的培训成为可能，将总部置于一个跨学科和协作的环境中。

Proteomics concerns the characterization of all proteins in a complex biological system, and can provide important insight into diseases, identifying biomarkers for diagnosis and personalized medicine. Today, protein characterization is carried out by a mass spectrometry experiment followed by data analysis with software. However, the data analysis has remained a principle bottleneck in the process. To deal with this challenge, I have developed innovative algorithms and analytical software routinely used by thousands of researchers in research labs worldwide, from both academy and industry. The long-term goal of the proposed research is to provide biological and health researchers with a powerful software platform for fully characterizing proteins with mass spectrometry. Backed with novel algorithms, the platform will be much improved in terms of accuracy and will have built-in quality control that guarantees high-quality results. Such a platform will enable and hasten the discovery process in many biological and health applications. Within this research proposal four short-term goals and subtasks will be pursued: (1) More accurate de novo sequencing: De novo sequencing identifies novel peptides that do not exist in any database. Our research will address the inadequate accuracy and lack of quality control in current methods. (2) New database search method with low false discovery rate (FDR): For peptides existing in a database, we will research new methods to identify more peptides while reducing FDR. (3) Post-translational modifications (PTM): PTMs and their quantity changes are very promising biomarkers. We will develop efficient computational methods to identify and quantify the modified peptides. (4) Human proteome profile: In order to better characterize recurring human peptides, we intend to build a human proteome profile by analyzing all publicly available human mass spectrometry data. Innovative algorithms are needed to achieve these goals. The research activities described in this proposal will enable effective training, placing HQP in a cross-disciplinary and collaborative environment.