Yang Mingyi, Ali Omer, Bjørås Magnar, Wang Junbai
Department of Microbiology, Oslo University Hospital and University of Oslo, Oslo, Norway.
Department of Medical Biochemistry, Oslo University Hospital and University of Oslo, Oslo, Norway.
iScience. 2023 Jul 3;26(8):107266. doi: 10.1016/j.isci.2023.107266. eCollection 2023 Aug 18.
Millions of single nucleotide variants (SNVs) exist in the human genome; however, it remains challenging to identify functional SNVs associated with diseases. We propose a non-encoding SNVs analysis tool bpb3, BayesPI-BAR version 3, aiming to identify the functional mutation blocks (FMBs) by integrating genome sequencing and transcriptome data. The identified FMBs display high frequency SNVs, significant changes in transcription factors (TFs) binding affinity and are nearby the regulatory regions of differentially expressed genes. A two-level Bayesian approach with a biophysical model for protein-DNA interactions is implemented, to compute TF-DNA binding affinity changes based on clustered position weight matrices (PWMs) from over 1700 TF-motifs. The epigenetic data, such as the DNA methylome can also be integrated to scan FMBs. By testing the datasets from follicular lymphoma and melanoma, bpb3 automatically and robustly identifies FMBs, demonstrating that bpb3 can provide insight into patho-mechanisms, and therapeutic targets from transcriptomic and genomic data.
人类基因组中存在数百万个单核苷酸变异（SNV）；然而，识别与疾病相关的功能性SNV仍然具有挑战性。我们提出了一种非编码SNV分析工具bpb3，即贝叶斯PI-BAR版本3，旨在通过整合基因组测序和转录组数据来识别功能突变块（FMB）。所识别的FMB显示出高频SNV、转录因子（TF）结合亲和力的显著变化，并且位于差异表达基因的调控区域附近。实施了一种两级贝叶斯方法以及蛋白质-DNA相互作用的生物物理模型，以基于来自1700多个TF基序的聚类位置权重矩阵（PWM）计算TF-DNA结合亲和力变化。表观遗传数据，如DNA甲基化组也可以整合到扫描FMB中。通过测试滤泡性淋巴瘤和黑色素瘤的数据集，bpb3能够自动且稳健地识别FMB，这表明bpb3可以从转录组和基因组数据中洞察病理机制和治疗靶点。
