identify single exon transcripts from cufflinks output

[xunong]

Hey guys,

Do you have an idea how to identify single exon transcripts from cufflinks output? Dose 'genes with only one exon' means exon_number "1" are eligible?
The transcripts.gtf file looks like:
chr1 Cufflinks transcript 11387 11823 1000 . . gene_id "CUFF.1"; transcript_id "CUFF.1.1"; FPKM "126581774.1320091337"; frac "1.000000"; conf_lo "23084341.949073"; conf_hi "230079206.314945"; cov "2.510922"; full_read_support "yes";
chr1 Cufflinks exon 11387 11823 1000 . . gene_id "CUFF.1"; transcript_id "CUFF.1.1"; exon_number "1"; FPKM "126581774.1320091337"; frac "1.000000"; conf_lo "23084341.949073"; conf_hi "230079206.314945"; cov "2.510922";
chr1 Cufflinks transcript 11874 29043 658 + . gene_id "CUFF.2"; transcript_id "CUFF.2.1"; FPKM "120296038.9757787585"; frac "0.143074"; conf_lo "104144214.907809"; conf_hi "136447863.043749"; cov "2.644528"; full_read_support "no";
chr1 Cufflinks exon 11874 12227 658 + . gene_id "CUFF.2"; transcript_id "CUFF.2.1"; exon_number "1"; FPKM "120296038.9757787585"; frac "0.143074"; conf_lo "104144214.907809"; conf_hi "136447863.043749"; cov "2.644528";
chr1 Cufflinks exon 12613 12721 658 + . gene_id "CUFF.2"; transcript_id "CUFF.2.1"; exon_number "2"; FPKM "120296038.9757787585"; frac "0.143074"; conf_lo "104144214.907809"; conf_hi "136447863.043749"; cov "2.644528";
chr1 Cufflinks exon 13221 29043 658 + . gene_id "CUFF.2"; transcript_id "CUFF.2.1"; exon_number "3"; FPKM "120296038.9757787585"; frac "0.143074"; conf_lo "104144214.907809"; conf_hi "136447863.043749"; cov "2.644528";
chr1 Cufflinks transcript 11874 14408 1 + . gene_id "CUFF.2"; transcript_id "NR_046018_1"; FPKM "0.0000000000"; frac "0.000000"; conf_lo "0.000000"; conf_hi "6894923.389408"; cov "0.000000"; full_read_support "no";

[dpryan]

Something like the following should work:

grep exon transcripts.gtf | awk '{print $10}' | sort | uniq -u

The will likely give the list of single-exon genes, assuming cufflinks is consistent in its file formatting.

Edit: I just saw that you wanted single-exon transcripts, not genes. Just change "print $10" to "print $12".