Wang Junmin
Data Sciences and Quantitative Biology, Discovery Sciences, Biopharmaceuticals R&D, AstraZeneca, Waltham, MA 02451, United States.
Bioinformatics. 2025 Jul 1;41(7). doi: 10.1093/bioinformatics/btaf353.
Understanding how small molecules perturb gene expression is critical for guiding drug discovery. We present DOSE-L1000-Viz, a Shiny application that facilitates comprehensive exploration of compound-induced transcriptomic responses across doses, time points, and cell types. Powered by a dose-response database, DOSE-L1000-Viz features interactive visualization, target-centric compound ranking based on efficacy and potency, and a signature search module using reference gene sets derived from generalized additive models. We benchmarked signatures derived from generalized additive models against traditional methods and demonstrated the utility of DOSE-L1000-Viz through use cases in transcription factor modulation and drug repurposing.
DOSE-L1000-Viz and the backend data are publicly accessible at: https://dosel1000.com. All code is publicly hosted on GitHub (https://github.com/JmWangBio/DOSEL1000Viz) and archived via Zenodo (https://doi.org/10.5281/zenodo.15532392).
了解小分子如何干扰基因表达对于指导药物发现至关重要。我们展示了DOSE-L1000-Viz，这是一个Shiny应用程序，有助于全面探索化合物在不同剂量、时间点和细胞类型下诱导的转录组反应。DOSE-L1000-Viz由剂量反应数据库提供支持，具有交互式可视化、基于疗效和效力的以靶点为中心的化合物排名，以及使用从广义相加模型派生的参考基因集的特征搜索模块。我们将从广义相加模型派生的特征与传统方法进行了基准测试，并通过转录因子调节和药物重新利用的用例展示了DOSE-L1000-Viz的实用性。
DOSE-L1000-Viz和后端数据可在以下网址公开获取：https://dosel1000.com。所有代码都在GitHub（https://github.com/JmWangBio/DOSEL1000Viz）上公开托管，并通过Zenodo（https://doi.org/10.5281/zenodo.15532392）存档。
