Long Non-Coding RNA Classification Using CNN and SVM

The system decides whether a transcript (cDNA or RNA sequence, 200-20 000 nt) is a long non-coding RNA (lncRNA) or a protein-coding mRNA. Two complementary models are combined: a support-vector machine with an RBF kernel on 384 hand-crafted sequence descriptors, and a one-dimensional convolutional neural network that reads the sequence itself. lncRNAs come from RNAcentral and protein-coding transcripts from GENCODE. Every result shows all model scores, an ORF map, the translated longest ORF and the CNN saliency along the transcript. A complete stand-alone MATLAB version generates the same sequence descriptors.

Category: Bioinformatics - Genomics and Sequence Analysis (Machine Learning, Deep Learning, CNN, SVM, MATLAB, Python)

Bioinformatics Project B08 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.