2
$\begingroup$

In the annotation gff3 provided for CHM13 v2, I noticed that for the SF3B3 gene there's no entry with type "gene" (third column).

However, transcripts and other annotations are still available. I checked on UCSC genome browser and transcripts still appear annotated:

UCSC screenshot

I was not able to find if this is associated with other aliases. I am trying to understand:

  • why the "gene" annotation is missing in the gff3
  • if this is temporary, i.e., the "gene" annotation will be updated later
$\endgroup$

1 Answer 1

2
$\begingroup$

Note that all of your transcripts for SF3B3 have a source_gene=ENSG00000189091.13 tag in the GFF3 file. A gene with that annotation does exist in the gff3 (see the grepped output below), though for some reason it has the identifier MSTRG.2232. That gene almost but not quite overlaps with the transcript SF3B3-201 and perfectly overlaps with the transcript CHM13_T0084548 represented in your browser shot.

# gene annotation
chr16   CAT     gene    76334966        76388859        .       +       .       source_gene_common_name=MSTRG.2232;source_gene=ENSG00000189091.13;gene_biotype=protein_coding;gene_id=CHM13_G0022147;gene_name=MSTRG.2232;transcript_modes=transMap,exRef;ID=CHM13_G0022147;Name=MSTRG.2232;source_transcript=N/A;alternative_source_transcripts=N/A;collapsed_gene_ids=N/A;collapsed_gene_names=N/A;paralogy=N/A;unfiltered_paralogy=N/A;alignment_id=N/A;frameshift=N/A;exon_anotation_support=N/A;intron_annotation_support=N/A;transcript_class=N/A;valid_start=N/A;valid_stop=N/A;proper_orf=N/A;extra_paralog=False
# transcript annotation
chr16   CAT transcript  76334997    76388857    10000   +   .   source_transcript=ENST00000302516.10;source_transcript_name=SF3B3-201;source_gene=ENSG00000189091.13;transcript_modes=transMap;gene_biotype=protein_coding;transcript_biotype=protein_coding;alignment_id=ENST00000302516.10-0;frameshift=nan;exon_annotation_support=1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1;intron_annotation_support=1,1,1,1,1,1,2,2,2,2,1,2,2,2,1,2,2,2,2,2,1,1,1,2,2;transcript_class=ortholog;valid_start=True;valid_stop=True;adj_start=76337842;adj_stop=76383007;proper_orf=True;level=2;protein_id=ENSP00000305790.5;transcript_support_level=1;hgnc_id=HGNC:10770;tag=basic,MANE_Select,appris_principal_1,CCDS;ccdsid=CCDS10894.1;havana_gene=OTTHUMG00000137582.8;havana_transcript=OTTHUMT00000268972.3;paralogy=nan;unfiltered_paralogy=nan;source_gene_common_name=SF3B3;transcript_id=CHM13_T0084551;gene_id=CHM13_G0022147;Parent=CHM13_G0022147;transcript_name=SF3B3-201;ID=CHM13_T0084551;Name=SF3B3;gene_name=SF3B3;alternative_source_transcripts=N/A;collapsed_gene_ids=N/A;collapsed_gene_names=N/A;extra_paralog=False
# transcript annotation
chr16   CAT     transcript      76334966        76388858        .       +       .       source_gene=ENSG00000189091.13;transcript_class=putative_novel_isoform;novel_5p_cap=True;novel_poly_a=True;transcript_biotype=protein_coding;gene_biotype=protein_coding;intron_rna_support=1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1;exon_rna_support=0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0;transcript_modes=exRef;exon_annotation_support=0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,0;intron_annotation_support=1,1,1,1,1,1,2,2,2,2,1,2,2,2,1,2,2,2,2,2,1,1,1,2,2,0;alignment_id=exRef-MSTRG.2232.2;source_gene_common_name=MSTRG.2232;transcript_id=CHM13_T0084548;gene_id=CHM13_G0022147;Parent=CHM13_G0022147;transcript_name=CHM13_T0084548;ID=CHM13_T0084548;Name=MSTRG.2232;gene_name=MSTRG.2232;source_transcript=N/A;alternative_source_transcripts=N/A;collapsed_gene_ids=N/A;collapsed_gene_names=N/A;paralogy=N/A;unfiltered_paralogy=N/A;frameshift=N/A;valid_start=N/A;valid_stop=N/A;proper_orf=N/A;extra_paralog=False

I'm not sure why those genes aren't tagged with the common name, and I'm not sure why it's not in your browser shot (whereas the alternate systematic gene NM_012426 is represented there; a quick grep suggests the NM_012426 IDs are not in the GFF3, and are interpolated by UCSC). But in general it's a good idea to always use the systematic identifier, because the common gene symbol/name won't always be trustworthy. If you use the systematic identifier, you should find the right gene.

Most likely explanation is a clash of competing methods, and a failure of the annotation method to hook the actual gene to the common name that you are looking for. It looks like some of the transcripts are longer than the gene, so maybe that's the reason. Possibly those transcripts derive from RNA-seq, and the gene annotator didn't find a good overlapping gene so it didn't get all the way linked. But I'm not enough of an annotation method expert to say for sure.

It looks like the annotation was done using CAT, so you could look there for more info. You could contact the author or post on the github if you want to learn more.

$\endgroup$
2
  • 1
    $\begingroup$ Thank you! Yes, the issue was reported on github. Your explanation is reasonable, should the Ensembl gene ID solve the issue, too? I tried to lift the coordinates from hg38 to chm13. After that I noticed that the sequence for SF3B3 (ENSG00000189091.13) is ~93% identical (through blast) between hg38 and chm13, so I believe that 7% difference might also have affected the annotation? $\endgroup$
    – gc5
    Mar 9, 2023 at 22:43
  • 1
    $\begingroup$ @gc5 Yeah, I would guess that something went weird with liftOff based on the difference, I would be curious to know if that 7% difference is due to real difference or different slices of highly similar sequence. But regardless I think the Ensembl gene ID should fix it. $\endgroup$ Mar 9, 2023 at 22:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .