Outlining a Strategy for Screening Non-coding RNAs on a Transcriptome Through Support Vector Machines

Abstract

Evidences that non-coding RNAs exert functions in organisms accumulate in the literature. Both computational predictions and experimental results have shown that, albeit not coding for a protein product, these transcripts play roles as diverse as catalytic activities and complex gene regulations, suggesting its therapeutic potential when applied to the study of pathogenic organisms. A target for such approach is the fungus Paracoccidioides brasiliensis (Pb), the ethyological agent of paracoccidioidomycosis, whose transcriptome has recently been elucidated. This work reports the compiling of a large training set and implementation of a framework of programs for sequence feature extraction, generating input for a Support Vector Machines algorithm for characterizing the coding potential of transcripts from a transcriptome.