UniDrug-Target: A Computational Tool to Identify Unique Drug Targets in Pathogenic Bacteria

BackgroundTargeting conserved proteins of bacteria through antibacterial medications has resulted in both the development of resistant strains and changes to human health by destroying beneficial microbes which eventually become breeding grounds for the evolution of resistances. Despite the availability of more than 800 genomes sequences, 430 pathways, 4743 enzymes, 9257 metabolic reactions and protein (three-dimensional) 3D structures in bacteria, no pathogen-specific computational drug target identification tool has been developed. MethodsA web server, UniDrug-Target, which combines bacterial biological information and computational methods to stringently identify pathogen-specific proteins as drug targets, has been designed. Besides predicting pathogen-specific proteins essentiality, chokepoint property, etc., three new algorithms were developed and implemented by using protein sequences, domains, structures, and metabolic reactions for construction of partial metabolic networks (PMNs), determination of conservation in critical residues, and variation analysis of residues forming similar cavities in proteins sequences. First, PMNs are constructed to determine the extent of disturbances in metabolite production by targeting a protein as drug target. Conservation of pathogen-specific protein's critical residues involved in cavity formation and biological function determined at domain-level with low-matching sequences. Last, variation analysis of residues forming similar cavities in proteins sequences from pathogenic versus non-pathogenic bacteria and humans is performed. ResultsThe server is capable of predicting drug targets for any sequenced pathogenic bacteria having fasta sequences and annotated information. The utility of UniDrug-Target server was demonstrated for Mycobacterium tuberculosis (H37Rv). The UniDrug-Target identified 265 mycobacteria pathogen-specific proteins, including 17 essential proteins which can be potential drug targets. Conclusions/SignificanceUniDrug-Target is expected to accelerate pathogen-specific drug targets identification which will increase their success and durability as drugs developed against them have less chance to develop resistances and adverse impact on environment. The server is freely available at http://117.211.115.67/UDT/main.html. The standalone application (source codes) is available at http://www.bioinformatics.org/ftp/pub/bioinfojuit/UDT.rar.背景：通过抗菌药物靶向细菌的保守蛋白，不仅催生了耐药菌株的演化，还会因破坏有益微生物而改变人体健康状态——这些有益微生物最终会成为耐药性进化的温床。尽管目前已积累超过800个细菌基因组序列、430条代谢通路、4743种酶、9257个代谢反应以及蛋白质三维（3D）结构数据，但尚未开发出针对病原体特异性的计算型药物靶点识别工具。 方法：本研究设计了一款名为UniDrug-Target的网络服务器，该工具整合细菌生物学信息与计算方法，以严格筛选病原体特异性蛋白作为药物靶点。除可预测病原体特异性蛋白的必需性、代谢瓶颈特性等属性外，本研究还开发并实现了三种全新算法，分别借助蛋白序列、结构域、三维结构以及代谢反应完成三类任务：其一，构建局部代谢网络（partial metabolic networks, PMNs），以评估将某蛋白作为药物靶点时对代谢物生成的干扰程度；其二，在序列匹配度较低的情况下，基于结构域水平鉴定参与空腔形成与生物学功能的病原体特异性蛋白关键残基的保守性；其三，对比致病菌、非致病菌与人类的蛋白序列中形成相似空腔的残基变异情况。 结果：该服务器可针对拥有FASTA序列与注释信息的任意已测序致病菌预测药物靶点。本研究以结核分枝杆菌（Mycobacterium tuberculosis）H37Rv为例，验证了UniDrug-Target服务器的实用性。UniDrug-Target共鉴定出265个分枝杆菌病原体特异性蛋白，其中包含17种可作为潜在药物靶点的必需蛋白。 结论与意义：UniDrug-Target有望加速病原体特异性药物靶点的识别进程，从而提升药物开发的成功率与持久性——因为针对此类靶点开发的药物更不易引发耐药性，且对环境的不良影响更小。该服务器可免费访问，网址为http://117.211.115.67/UDT/main.html。其独立应用程序（源代码）可从http://www.bioinformatics.org/ftp/pub/bioinfojuit/UDT.rar 获取。