Identification of long non-coding transcripts with feature selection: a comparative study