SBIR Phase I: Automated Humanization of Antibodies Directed to Protein and Carbohydrate Antigens

这个小企业创新研究（SBIR）项目的更广泛的影响/商业潜力将是开发一个在线的、全自动的平台，用于设计高亲和力抗体，作为潜在的候选药物。基于抗体的药物的成功引起了人们对更快、更有效的方法来发现和优化抗体的兴趣。该项目的目标将是开发和实施一种计算方法，用于生产人源化抗体的蛋白质序列。这将通过提供软件来实现，该软件允许科学家将他们的一些初始实验转移到云端，并通过使用计算方法更快地获得结果。用这种方法代替的典型实验大约需要一个月的时间，花费数万美元。在这种方法的帮助下，这个时间可以减少到点击一个按钮。预期的新药开发时间和成本的节省将改善候选药物的特性，提高临床研究的成功率，加速新药的商业化。这个SBIR一期项目提出开发一个抗体设计平台，最大限度地提高人类含量，保留亲和力，增强可开发性，并产生知识产权。在开发基于抗体的药物时，免疫原性是一个关键问题。人源化的目的是增加从非人类来源（如小鼠或兔杂交瘤）获得的抗体的人源含量。目标是简化人源化抗体的设计，目前这涉及许多劳动密集型步骤。该策略是为客户提供在线预测服务，精确设计并专注于抗体分析。其次，目标是解决针对蛋白质和碳水化合物抗原的人源化抗体的新需求。目前的预测软件偏向于对蛋白质抗原特异性的训练系列。该计划是利用从蛋白质和碳水化合物靶点中获得的抗伯克氏菌小鼠单克隆候选抗体池。这将导致有能力评估正在开发的培训系列的效用，以执行针对这两类抗原的抗体的人源化，并利用所获得的信息进一步优化预测程序。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) project will be to develop an online, fully automated platform for designing high-affinity antibodies for use as potential drug candidates. The success of antibody-based drugs has generated interest in faster and more efficient methods to discover and optimize antibodies. The goal of this project will be to develop and implement a computation method for producing protein sequences of humanized antibodies. This will be achieved by providing software that allows scientists to move some of their initial experiments into the cloud, and achieve results much more quickly by using computational methods. The typical experiments that will be replaced by this method take about a month of time and cost tens of thousands of dollars. With the aid of such a method, this time can be cut back to the click of a button. The expected savings in time and cost for new drug development will improve the features of drug candidates, enhancing the success rate of clinical studies, and accelerating the commercialization of new drugs.This SBIR Phase I project proposes the development of a platform for antibody design that maximizes human content, retention of affinity, enhancement of developability, and generation of intellectual property. Immunogenicity is a critical concern when developing an antibody-based drug. Humanization is designed to increase the human content of an antibody originally obtained from sources other than human, such as mouse or rabbit hybridomas. The objective is to simplify the design of humanized antibodies, which currently involves many labor-intensive steps. The strategy is to offer clients online prediction services precisely designed and focused on antibody analysis. Secondly, the goal is to address the emerging need for humanizing antibodies directed against both protein and carbohydrate antigens. Current prediction software is biased towards training series that are specific for protein antigens. The plan is to use a pool of anti-Burkholderia mouse monoclonal antibody candidates have been derived from protein and carbohydrate targets. This will result in the ability to assess the utility of the training series under development to perform humanization of antibodies directed against both classes of antigens, and use the information obtained to further optimize prediction procedures.