COUSCOus: improved protein contact prediction using an empirical Bayes covariance estimator

BMC Bioinformatics

Research Article

Reda Rawi¹  Khalid Kunji¹  Ehsan Ullah¹  Michael Aupetit¹  Raghvendra Mall¹  Halima Bensmail¹  Mohammed El Anbari² 

[1] Computational Science and Engineering, Qatar Computing Research Institute, Hamad Bin Khalifa University, Doha, Qatar;Division of Biomedical Informatics, Sidra Medical and Research Center, Doha, Qatar;

关键词: Residue-residue contact prediction;    Shrinkage;    GLasso;   

DOI  :  10.1186/s12859-016-1400-3

 received in 2016-06-17, accepted in 2016-12-01,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundThe post-genomic era with its wealth of sequences gave rise to a broad range of protein residue-residue contact detecting methods. Although various coevolution methods such as PSICOV, DCA and plmDCA provide correct contact predictions, they do not completely overlap. Hence, new approaches and improvements of existing methods are needed to motivate further development and progress in the field. We present a new contact detecting method, COUSCOus, by combining the best shrinkage approach, the empirical Bayes covariance estimator and GLasso.ResultsUsing the original PSICOV benchmark dataset, COUSCOus achieves mean accuracies of 0.74, 0.62 and 0.55 for the top L/10 predicted long, medium and short range contacts, respectively. In addition, COUSCOus attains mean areas under the precision-recall curves of 0.25, 0.29 and 0.30 for long, medium and short contacts and outperforms PSICOV. We also observed that COUSCOus outperforms PSICOV w.r.t. Matthew’s correlation coefficient criterion on full list of residue contacts. Furthermore, COUSCOus achieves on average 10% more gain in prediction accuracy compared to PSICOV on an independent test set composed of CASP11 protein targets. Finally, we showed that when using a simple random forest meta-classifier, by combining contact detecting techniques and sequence derived features, PSICOV predictions should be replaced by the more accurate COUSCOus predictions.ConclusionWe conclude that the consideration of superior covariance shrinkage approaches will boost several research fields that apply the GLasso procedure, amongst the presented one of residue-residue contact prediction as well as fields such as gene network reconstruction.

【 授权许可】

CC BY   
© The Author(s) 2016

【 图 表 】

12864_2016_2793_Article_IEq40.gif

12864_2017_3683_Article_IEq2.gif

12864_2017_3492_Article_IEq1.gif

12864_2016_2913_Article_IEq39.gif

12864_2015_2055_Article_IEq34.gif

12864_2017_4133_Article_IEq18.gif

12864_2017_3487_Article_IEq38.gif

12864_2017_4132_Article_IEq30.gif

12864_2017_4132_Article_IEq28.gif