De Novo Approach to Classify Protein-Coding and Noncoding Transcripts Based on Sequence Composition

Each day, more and more transcripts are being discovered along the genome (especially in poorly annotated species) thanks to the rapid progress of high-throughput technology such as RNA sequencing. However, this situation unravels the challenge of how to classify the newly identified transcripts into protein coding or noncoding. Here, we describe a de novo approach named coding–noncoding index (CNCI), a powerful signature tool by profiling adjoining nucleotide triplets (ANT) to effectively distinguish between protein-coding and noncoding sequences independently of known annotations. The main advantage of CNCI is its ability to accurately classify transcripts assembled from whole-transcriptome sequencing data in a cross-species manner, which allowed it to be used for all vertebrates and invertebrates based on the training data of well-annotated species (such as human and Arabidopsis). In this chapter, we illustrate the CNCI method in detail through an example of RNA-sequencing data generated from six biological replicates of six mouse tissues. CNCI software is available at http://www.bioinfo.org/software/cnci .