首页> 外文期刊>Chemometrics and Intelligent Laboratory Systems >iRSpot-SPI: Deep learning-based recombination spots prediction by incorporating secondary sequence information coupled with physio-chemical properties via Chou's 5-step rule and pseudo components
【24h】

iRSpot-SPI: Deep learning-based recombination spots prediction by incorporating secondary sequence information coupled with physio-chemical properties via Chou's 5-step rule and pseudo components

机译:IRSPOT-SPI:基于深度学习的重组点通过Chou的5步规则和伪组分结合与物理化学性质的二次序列信息进行预测

获取原文
获取原文并翻译 | 示例

摘要

Meiotic recombination plays an important role in the process of genetic diversity generation. "Hotspots" are regions that show a higher rate of meiotic recombination, while the regions with a lower rate are called "cold spots". It has a great effect on the genome evolution via gene conversion or mutagenesis. According to recent research, recombination is present in uneven distribution across the genome. Many computational methods have been developed using secondary sequence information or physiochemical properties of nucleotide descriptor for the prediction of hotspots and cold spots, which are computationally cheap and fast in performance rather than web-lab experiments, but the correlations between nucleotides pairs at different positions along DNA sequence is often ignored, which conceal a very important predictive information. In this study, we have proposed a deep neural network to predict recombination spots by fusing both the secondary sequence information and physiochemical derived features. Our deep learning algorithm leverage's deep dense architecture by showing its effectiveness over the state-of-the-art methods with a classification accuracy of 90.04%, sensitivity of 92.21%, specificity of 92.11% and area under the curve of 0.9801. Moreover, it is anticipated, that our model will provide novel insight into basic research, drug designing, academic research and recombination spots studies particularly. All the methodology and python-based source code is publicly available for the users at https://github.com/zaheerkhancs/irSpot_SPI along with publicly accessible web server using the proposed predictor.
机译:减数分裂重组在遗传多样性的过程中起着重要作用。 “热点”是具有更高的减数率重组率的地区,而具有较低速率的区域称为“冷点”。通过基因转化或诱变对基因组进化产生了很大的影响。根据最近的研究,重组存在于基因组的不均匀分布中。使用核苷酸描述符的二次序列信息或生理化学特性来开发许多计算方法,用于预测热点和冷点,这些热点和寒冷的斑点是在性能而不是网上实验室实验中计算的,但核苷酸对不同位置的相关性DNA序列通常被忽略,隐藏了一个非常重要的预测信息。在这项研究中,我们提出了一种深度神经网络,通过融合二次序列信息和生理化学衍生特征来预测重组斑点。我们深入的学习算法利用了对最先进的方法的有效性来利用了90.04%,灵敏度为92.21%,92.11%的特异性和0.9801的面积的效果来杠杆化。此外,预计,我们的模型将提供对基础研究,药物设计,学术研究和重组斑点的新颖洞察力。所有的方法和基于Python的源代码都可以公开为Https://github.com/zaheerkhancs/irspot_spi的用户公开使用,并使用所提出的预测器与公开访问的Web服务器一起使用。

著录项

相似文献

  • 外文文献
  • 中文文献
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号