A semi-supervised deep learning approach for predicting the functional effects of genomic non-coding variations.
A semi-supervised deep learning approach for predicting the functional effects of genomic non-coding variations.
复制标题
DOI：
10.1186/s12859-021-03999-8
复制
发表时间：
2021-06-02
影响因子：
3
通讯作者：
Nakai K
中科院分区：
文献类型：
--
作者：
Jia H;Park SJ;Nakai K
了解非编码变体的功能效应非常重要，因为它们通常与基因表达改变和疾病发展相关。在过去的几年里，已经开发了许多计算工具来预测其功能影响。然而，在处理数据的稀缺性的内在困难导致的必要性，以进一步改善算法。在这项工作中，我们提出了一种新的方法，采用带有伪标签的半监督深度学习模型，该模型利用了从实验注释和未注释数据中学习的优势。我们在GM 12878、HepG 2和K562细胞系中制备了具有组蛋白标记、DNA可及性和序列背景的已知功能性非编码变体。将我们的方法应用到数据集上，与现有的工具相比，表现出了出色的性能。我们的研究结果还表明，与没有伪标签的监督模型相比，具有伪标签的半监督模型具有更高的预测性能。有趣的是，用特定细胞系中的数据训练的模型不太可能在其他细胞系中成功，这意味着非编码变体的细胞类型特异性。值得注意的是，我们发现，DNA的可及性显着有助于变异的功能后果，这表明开放的染色质构象的重要性，建立非编码变异与基因调控的相互作用之前。结合伪标记的半监督深度学习模型在使用有限数据集进行学习方面具有优势，这在生物学中并不罕见。我们的研究提供了一种有效的方法来寻找可能与各种生物现象，包括人类疾病相关的非编码突变。在线版本包含补充材料，可通过10.1186/s12859-021-03999-8获得。
Understanding the functional effects of non-coding variants is important as they are often associated with gene-expression alteration and disease development. Over the past few years, many computational tools have been developed to predict their functional impact. However, the intrinsic difficulty in dealing with the scarcity of data leads to the necessity to further improve the algorithms. In this work, we propose a novel method, employing a semi-supervised deep-learning model with pseudo labels, which takes advantage of learning from both experimentally annotated and unannotated data. We prepared known functional non-coding variants with histone marks, DNA accessibility, and sequence context in GM12878, HepG2, and K562 cell lines. Applying our method to the dataset demonstrated its outstanding performance, compared with that of existing tools. Our results also indicated that the semi-supervised model with pseudo labels achieves higher predictive performance than the supervised model without pseudo labels. Interestingly, a model trained with the data in a certain cell line is unlikely to succeed in other cell lines, which implies the cell-type-specific nature of the non-coding variants. Remarkably, we found that DNA accessibility significantly contributes to the functional consequence of variants, which suggests the importance of open chromatin conformation prior to establishing the interaction of non-coding variants with gene regulation. The semi-supervised deep learning model coupled with pseudo labeling has advantages in studying with limited datasets, which is not unusual in biology. Our study provides an effective approach in finding non-coding mutations potentially associated with various biological phenomena, including human diseases. The online version contains supplementary material available at 10.1186/s12859-021-03999-8.
登录
查看更多内容
影响因子：
30.8
作者：
Lee D;Gorkin DU;Baker M;Strober BJ;Asoni AL;McCallion AS;Beer MA
通讯作者：
Beer MA
影响因子：
30.8
作者：
Kircher, Martin;Witten, Daniela M.;Jain, Preti;O'Roak, Brian J.;Cooper, Gregory M.;Shendure, Jay
通讯作者：
Shendure, Jay
影响因子：
30.8
作者：
Ionita-Laza I;McCallum K;Xu B;Buxbaum JD
通讯作者：
Buxbaum JD
DOI：
10.1073/pnas.0903103106
发表时间：
2009-06-09
影响因子：
11.1
作者：
Hindorff, Lucia A.;Sethupathy, Praveen;Manolio, Teri A.
通讯作者：
Manolio, Teri A.
影响因子：
46.9
作者：
Bernstein, Bradley E.;Stamatoyannopoulos, John A.;Costello, Joseph F.;Ren, Bing;Milosavljevic, Aleksandar;Meissner, Alexander;Kellis, Manolis;Marra, Marco A.;Beaudet, Arthur L.;Ecker, Joseph R.;Farnham, Peggy J.;Hirst, Martin;Lander, Eric S.;Mikkelsen, Tarjei S.;Thomson, James A.
通讯作者：
Thomson, James A.