why exon_number 0 existed/generated in gtf ?

[alejandrogzi]

why exon_number 0 existed/generated in gtf ?

[HUNNNGRY]

Yes, you are right, I've checked ENST00000468844.1 and ENST00000477403.1, both of which has no CDS annotaion in genocode.v27.gtf, only transcript and exons.
Currently everything works fine, many thanks for your contribution : )

[alejandrogzi]

Hi @HUNNNGRY ,

Thanks for using bed2gtf and for reporting this bug! I ran some tests here and confirm that this is a bug due to the some error within an iteration. I have fix this, now if you convert your test.bed12 file, it should look like this:

#provider: bed2gtf
#version: 1.8.0
#contact: github.com/alejandrogzi/bed2gtf
#date: 2024-1-3
chrX	bed2gtf	gene	119786504	119791643	.	-	.	gene_id "ENSG00000198918.7";
chrX	bed2gtf	transcript	119786504	119787456	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1";
chrX	bed2gtf	exon	119787281	119787456	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "1"; exon_id "ENST00000477403.1.1";
chrX	bed2gtf	exon	119786504	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "2"; exon_id "ENST00000477403.1.2";
chrX	bed2gtf	transcript	119786506	119791643	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	exon	119791574	119791643	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	CDS	119791574	119791576	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	start_codon	119791574	119791576	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	exon	119789908	119790011	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "2"; exon_id "ENST00000361575.3.2";
chrX	bed2gtf	CDS	119789908	119790011	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "2"; exon_id "ENST00000361575.3.2";
chrX	bed2gtf	exon	119786506	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	CDS	119786687	119786732	.	-	1	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	stop_codon	119786684	119786686	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	five_prime_utr	119791577	119791643	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	three_prime_utr	119786506	119786683	.	-	1	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	transcript	119786506	119791595	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1";
chrX	bed2gtf	exon	119789908	119791595	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1"; exon_number "1"; exon_id "ENST00000468844.1.1";
chrX	bed2gtf	exon	119786506	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1"; exon_number "2"; exon_id "ENST00000468844.1.2";

Your catch will force a new release of bed2gtf; however, the remaining improvements planned will take 1-2 days because of my responsabilities, delaying the new release by that amount of time. Sorry for any inconvenience.

You can still run the fixed bed2gtf version by:

git clone https://github.com/alejandrogzi/bed2gtf.git && cd bed2gtf

and

cargo run release -- --bed test.bed12 --isoforms test.isoform --output test.gtf

Please let me know if everything works as intended and do not hesitate to communicate another bug/problem. I will let you know when the new release is published (probably tomorrow), otherwise you could star this repo and the notifications will appear automatically in your home page.

Best,
Alejandro

[HUNNNGRY]

Thank you for your quick reply, the new version seem output correct result for test data.
But I'm wondering why only one CDS/Start_codon/UTR is generated for one gene with multiple transcripts ? I think its common to output all CDS/Start_codon/UTR under each transcripts of a gene (like gencode.v27.gtf downloaded from GENCODE project).
And how did you select the representative transcript for each gene for this version ? (maybe the transcript with longest CDS ? or just random ?)

[alejandrogzi]

@HUNNNGRY,

I am glad to hear that!

But I'm wondering why only one CDS/Start_codon/UTR is generated for one gene with multiple transcripts ? I think its common to output all CDS/Start_codon/UTR under each transcripts of a gene [...] And how did you select the representative transcript for each gene for this version ? (maybe the transcript with longest CDS ? or just random ?)

This happens because of the coordinates in your .bed file. bed2gtf does not choose the longest CDS, the annotation of features is transcript-dependent. Here is the step-by-step explanation:

Your .bed file looks like this:

chrX	119786505	119791643	ENST00000361575.3	0	-	119786683	119791576	0	3	227,104,70,	0,3402,5068,
chrX	119786505	119791595	ENST00000468844.1	0	-	119791595	119791595	0	2	227,1688,	0,3402,
chrX	119786503	119787456	ENST00000477403.1	0	-	119787456	119787456	0	2	229,176,	0,777,

Note that for ENST00000468844.1 and ENST00000477403.1 the CDS start and end are equal:

ENST00000468844.1 [...] 119791595	119791595
ENST00000477403.1 [...] 119787456	119787456

This will not allow bed2gtf to move around exon boundaries to establish frame-corrected CDSs. When those CDS are corrected, for example:

from

ENST00000477403.1 [...] 119787456	119787456

to

ENST00000477403.1 [...] 119786503	119787456

and bed2gtf is run, the output is now complete for that transcript:

#provider: bed2gtf
#version: 1.8.0
#contact: github.com/alejandrogzi/bed2gtf
#date: 2024-1-3
chrX	bed2gtf	gene	119786504	119791643	.	-	.	gene_id "ENSG00000198918.7";
chrX	bed2gtf	transcript	119786504	119787456	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1";
chrX	bed2gtf	exon	119787281	119787456	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "1"; exon_id "ENST00000477403.1.1";
chrX	bed2gtf	CDS	119787281	119787456	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "1"; exon_id "ENST00000477403.1.1";
chrX	bed2gtf	start_codon	119787454	119787456	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "1"; exon_id "ENST00000477403.1.1";
chrX	bed2gtf	exon	119786504	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "2"; exon_id "ENST00000477403.1.2";
chrX	bed2gtf	CDS	119786507	119786732	.	-	1	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "2"; exon_id "ENST00000477403.1.2";
chrX	bed2gtf	stop_codon	119786504	119786506	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000477403.1"; exon_number "2"; exon_id "ENST00000477403.1.2";
chrX	bed2gtf	transcript	119786506	119791643	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	exon	119791574	119791643	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	CDS	119791574	119791576	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	start_codon	119791574	119791576	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "1"; exon_id "ENST00000361575.3.1";
chrX	bed2gtf	exon	119789908	119790011	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "2"; exon_id "ENST00000361575.3.2";
chrX	bed2gtf	CDS	119789908	119790011	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "2"; exon_id "ENST00000361575.3.2";
chrX	bed2gtf	exon	119786506	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	CDS	119786687	119786732	.	-	1	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	stop_codon	119786684	119786686	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3"; exon_number "3"; exon_id "ENST00000361575.3.3";
chrX	bed2gtf	five_prime_utr	119791577	119791643	.	-	0	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	three_prime_utr	119786506	119786683	.	-	1	gene_id "ENSG00000198918.7"; transcript_id "ENST00000361575.3";
chrX	bed2gtf	transcript	119786506	119791595	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1";
chrX	bed2gtf	exon	119789908	119791595	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1"; exon_number "1"; exon_id "ENST00000468844.1.1";
chrX	bed2gtf	exon	119786506	119786732	.	-	.	gene_id "ENSG00000198918.7"; transcript_id "ENST00000468844.1"; exon_number "2"; exon_id "ENST00000468844.1.2";

Note now that the only transcript that lacks a CDS is ENSG00000198918.7. All these results are consistent with the counterpart tools from USCS's utils (bedToGenePred | genePredToGtf (since bed2gtf uses their core functionatlity).

Hope this helps!

Best,
Alejandro