Summary: Protein quality assessment is a long-standing problem inbioinformatics. For more than a decade we have developed state-of-artpredictors by carefully selecting and optimising inputs to a machine learningmethod. The correlation has increased from 0.60 in ProQ to 0.81 in ProQ2 and0.85 in ProQ3 mainly by adding a large set of carefully tuned descriptions of aprotein. Here, we show that a substantial improvement can be obtained usingexactly the same inputs as in ProQ2 or ProQ3 but replacing the support vectormachine by a deep neural network. This improves the Pearson correlation to 0.90(0.85 using ProQ2 input features). Availability: ProQ3D is freely available both as a webserver and astand-alone program at http://proq3.bioinfo.se/
展开▼
