Prediction of the Structure of Therapeutic Antibodies with their Antigens

用其抗原预测治疗性抗体的结构 项目总结 抗体在识别外来入侵者方面发挥着关键作用。由于它们的高度亲和力和特异性，它们 已被用作治疗分子和用于传感和组装的生物技术组件。 最近的高通量测序和纳米流体技术已经阐明了大集合(103-104) 天真和暴露于抗原的抗体序列，现在可以确定一套完整的病毒 个人根据自身抗体所遇到的情况。除了他们的生物学、医学和 技术重要性，关于抗体的广泛知识使其成为理想的模型系统 研究蛋白质结合和识别。 研究蛋白质结合和识别的可靠工具包是完全解锁丰富蛋白质的缺失环节 抗体和抗原库中的信息。先前的工作证明预测抗体是成功的 结构，这项提议集中在对接问题上。而对接算法在本地是可靠的 搜索和小的构象变化，在寻找大的抗原到 识别表位，并在有主干灵活性或 同源模型起始结构中的不确定性。考虑绑定诱导的主干 构象变化仍然是蛋白质-蛋白质对接领域的中心困难，主要是由于 抽样限制。另一个挑战是许多病毒外壳和细菌蛋白是糖基化的。 葡聚糖具有良好的水合性和柔韧性；这些修饰通常通过对接完全忽略 算法。 这项研究的长期目标是准确预测抗体和抗体-抗原的结构。 因此，它们可用于破译生物机制和设计改进的治疗方法。 因此，当前项目的前两个目标是(1)发展快速、积极、灵活的骨干网对接 方法，以及(2)扩展对接以包括糖基化抗原。最后，第三个目标将是(3)应用 抗体建模和对接以确定乳糜泻和肺疾病的生物标志物和治疗方法 高血压。

Prediction of the Structure of Therapeutic Antibodies with their Antigens PROJECT SUMMARY Antibodies play a critical role for recognition of foreign intruders. Due to their high affinity and specificity, they have been exploited as therapeutic molecules and biotechnological components for sensing and assembly. Recent high-throughput sequencing and nanofluidics technologies have elucidated large sets (103–104) of naïve and antigen-exposed antibody sequences, and it is now possible to determine a complete set of viruses that an individual has encountered based on one’s antibodies. In addition to their biological, medical, and technological importance, the extensive knowledge about antibodies makes them an ideal model system for studying protein binding and recognition. A reliable toolkit to study protein binding and recognition is the missing link to fully unlock the bountiful information in antibody and antigen repertoires. Prior work demonstrated success in predicting antibody structures, and this proposal focuses on the docking problem. While docking algorithms are reliable for local searches and small conformational changes, significant challenges remain in searching large antigens to identify epitopes and in determining the correct binding orientation when there is backbone flexibility or uncertainty in the homology-modeled starting structures. Accounting for binding-induced backbone conformational changes remains the central difficulty in the protein–protein docking field, primarily due to sampling limitations. An additional challenge is that many viral coat and bacterial proteins are glycosylated. Glycans are well hydrated and can be flexible; these modifications are typically ignored entirely by docking algorithms. The long-term goal of this research is the accurate prediction of structures of antibodies and antibody–antigen complexes such that they are useful to decode biological mechanisms and engineer improved therapeutics. Thus, the first two aims of the current project are to (1) develop fast, aggressive, flexible backbone docking approaches, and (2) extend docking to include glycosylated antigens. Finally, the third aim will be to (3) apply antibody modeling and docking to determine biomarkers and therapeutics for celiac disease and pulmonary hypertension.