Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.
Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.
复制标题
DOI：
10.1093/bioinformatics/bty575
复制
发表时间：
2018-09-01
期刊：
影响因子：
--
通讯作者：
Kundaje A
中科院分区：
文献类型：
--
作者：
Greenside P;Shimko T;Fordyce P;Kundaje A
转录因子以组合方式结合调控DNA序列来调节基因表达。深度神经网络 (DNN) 可以学习与转录因子结合和染色质可及性相关的调控 DNA 序列中编码的顺式调控语法。已经开发了几种特征归因方法，用于估计任何输入 DNA 序列中的单个特征（核苷酸或基序）对其 DNN 模型的相关输出预测的预测重要性。然而，这些方法并没有揭示模型编码的高阶特征交互。我们提出了一种称为深度特征交互图 (DFIM) 的新方法，可以有效估计任何输入 DNA 序列中所有特征对之间的交互。 DFIM 准确识别嵌入模拟调控 DNA 序列中的真实基序相互作用。 DFIM 从体内 TF 结合模型中识别出 GATA1 和 TAL1 基序之间的协同相互作用。 DFIM 通过体外 TF 结合模型揭示了酵母中 Cbf1 TF 核心基序侧翼核苷酸的上位相互作用。我们还将 DFIM 应用于体内染色质可及性的调控序列模型，以揭示调控遗传变异与目标 TF 的近端基序之间的相互作用，并通过 TF 结合数量性状位点进行验证。我们的方法在提高基因组学深度学习模型的可解释性方面取得了重大进展。代码位于：https://github.com/kundajelab/dfim。 补充数据可在生物信息学在线获取。
Transcription factors bind regulatory DNA sequences in a combinatorial manner to modulate gene expression. Deep neural networks (DNNs) can learn the cis-regulatory grammars encoded in regulatory DNA sequences associated with transcription factor binding and chromatin accessibility. Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output prediction from a DNN model. However, these methods do not reveal higher-order feature interactions encoded by the models. We present a new method called Deep Feature Interaction Maps (DFIM) to efficiently estimate interactions between all pairs of features in any input DNA sequence. DFIM accurately identifies ground truth motif interactions embedded in simulated regulatory DNA sequences. DFIM identifies synergistic interactions between GATA1 and TAL1 motifs from in vivo TF binding models. DFIM reveals epistatic interactions involving nucleotides flanking the core motif of the Cbf1 TF in yeast from in vitro TF binding models. We also apply DFIM to regulatory sequence models of in vivo chromatin accessibility to reveal interactions between regulatory genetic variants and proximal motifs of target TFs as validated by TF binding quantitative trait loci. Our approach makes significant strides in improving the interpretability of deep learning models for genomics. Code is available at: https://github.com/kundajelab/dfim. Supplementary data are available at Bioinformatics online.
登录
查看更多内容
影响因子：
14.9
作者：
Kheradpour P;Kellis M
通讯作者：
Kellis M
影响因子：
30.8
作者：
Corces, M. Ryan;Buenrostro, Jason D.;Wu, Beijing;Greenside, Peyton G.;Chan, Steven M.;Koenig, Julie L.;Snyder, Michael P.;Pritchard, Jonathan K.;Kundaje, Anshul;Gkeenleaf, William J.;Majeti, Ravindra;Chang, Howard Y.
通讯作者：
Chang, Howard Y.
合并的芯片seq连接转录因子与复杂疾病风险结合的变化。
影响因子：
64.5
作者：
Tehranchi AK;Myrthil M;Martin T;Hie BL;Golan D;Fraser HB
通讯作者：
Fraser HB
影响因子：
4.5
作者：
Friedman, JH
通讯作者：
Friedman, JH
影响因子：
64.8
作者：
Gerstein, Mark B.;Kundaje, Anshul;Hariharan, Manoj;Landt, Stephen G.;Yan, Koon-Kiu;Cheng, Chao;Mu, Xinmeng Jasmine;Khurana, Ekta;Rozowsky, Joel;Alexander, Roger;Min, Renqiang;Alves, Pedro;Abyzov, Alexej;Addleman, Nick;Bhardwaj, Nitin;Boyle, Alan P.;Cayting, Philip;Charos, Alexandra;Chen, David Z.;Cheng, Yong;Clarke, Declan;Eastman, Catharine;Euskirchen, Ghia;Frietze, Seth;Fu, Yao;Gertz, Jason;Grubert, Fabian;Harmanci, Arif;Jain, Preti;Kasowski, Maya;Lacroute, Phil;Leng, Jing;Lian, Jin;Monahan, Hannah;O'Geen, Henriette;Ouyang, Zhengqing;Partridge, E. Christopher;Patacsil, Dorrelyn;Pauli, Florencia;Raha, Debasish;Ramirez, Lucia;Reddy, Timothy E.;Reed, Brian;Shi, Minyi;Slifer, Teri;Wang, Jing;Wu, Linfeng;Yang, Xinqiong;Yip, Kevin Y.;Zilberman-Schapira, Gili;Batzoglou, Serafim;Sidow, Arend;Farnham, Peggy J.;Myers, Richard M.;Weissman, Sherman M.;Snyder, Michael
通讯作者：
Snyder, Michael