Support for New Bioinformatics Methods Development

支持新的生物信息学方法开发

• 批准号: 10596966
• 负责人: RUCHIR SHAH
• 金额: $ 30万
• 依托单位: SCIOME, LLC
• 依托单位国家: 美国
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-03-24 至 2022-09-23
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10596966
• 关键词: Bioinformatics; Classification; Collection; Computers; Contractor; Contracts; Data Analyses; Databases; Development; Hepatocyte; Image; Image Analysis; In Vitro; Infrastructure; Java; Libraries; Maintenance; Manuscripts; Methods; National Institute of Environmental Health Sciences; Peer Review; Programming Languages; Resources; Specific qualifier value; Text; bioinformatics tool; cheminformatics; cytotoxicity; deep learning; glyphosate; high throughput screening; innovation; method development; tool

New bioinformatics method development support includes a variety of different DNTP projects. 1) BMDExpress 3 is a new version supported by the contractor that will offer IVIVE analysis using the HTTK library developed by EPA and others. Sciome ported the HTTK library from R to Java so that BMDExpress 3 can access the methods to perform the calculations. 2) Automated image classification of in vitro cytotoxicity programming was used for deep learning of image analysis data on HepaRG and hepatocytes in high throughput screening studies. Results indicated successful classification with two-class and multi-class classification accuracies. A large collection of images was classified using this framework for the glyphosate studies and a peer-reviewed manuscript is in progress. 3) for OHAT/DNTP evidence mapping was provided by implementing a ‘clean-up’ tool in SWIFT-Review that allows users to remove tags for which the percentage if below a minimum threshold. A tool has also integrated the option to have full-text in SWIFT-Review.

新的生物信息学方法开发支持包括各种不同的DNTP项目。 1)BMDExpress 3是承包商支持的新版本，将使用EPA和其他机构开发的HTTK库提供IVIVE分析。 Sciome将HTTK库从R移植到Java，以便BMDExpress 3可以访问执行计算的方法。2)体外细胞毒性编程的自动图像分类用于高通量筛选研究中HepaRG和肝细胞图像分析数据的深度学习。结果表明，成功的分类与两类和多类分类精度。 大量的图像收集使用这个框架进行分类，用于草甘膦研究，同行评审的手稿正在进行中。3)OHAT/DNTP的证据映射是通过在SWIFT-Review中实施“清理”工具来提供的，该工具允许用户删除百分比低于最低阈值的标签。一个工具还集成了在SWIFT-Review中提供全文的选项。