knownCanonical

[Alex Lenail]

Hello, 

I'd like to download a set of canonical splice isoforms for hg38 (gencode 29), with a single splice isoform per gene symbol. Right now, the knownCanonical table available for download contains 64,792 rows, which, when merged with kgXref, reveals multiple "canonical" isoforms per gene symbol for many genes. The FAQ states: 

I just want to download a gene set with a single entry per gene. Where can I find this?

We have data tables named knownCanonical available for different assemblies comprised of a single transcript/isoform per gene.

...

For hg38, the knownCanonical table is a subset of the GENCODE v29 track. As opposed to the hg19 equivalent which generally used the longest isoform for indentification, this table is defined as follows:

knownCanonical identifies the canonical isoform of each cluster ID or gene using the ENSEMBL gene IDs to define each cluster. The canonical transcript is chosen using the APPRIS principal transcript when available. If no APPRIS tag exists for any transcript associated with the cluster, then a transcript in the BASIC set is chosen. If no BASIC transcript exists, then the longest isoform is used.

It can be downloaded directly from the hg38 downloads database or by using the Table Browser.

It seems there's likely a canonical isoform for each geneSymbol / gene cluster combination. I'm not aware of what a gene cluster is; the table schema at UCSC just tells me it's 

Which cluster of transcripts this belongs to in knownIsoforms

Would you be willing to point me towards documentation as to what those clusterIds are, or better yet, how I can get a table of unique canonical isoforms per gene symbol? 

Many thanks, 

--Alex

[Alex Lenail]

And a follow-up: 

I had anticipated that the ENST* names would be unique in the knownGene annotation file, but I see many rows like the following: 

ENST00000313871.8 chrX + 1591592 1602514 1593462 1601594 5 1591592,1593443,1595383,1599191,1600658, 1591769,1594224,1595532,1599432,1602514, Q02040 uc004fpn.4

ENST00000313871.8 chrY + 1591592 1602514 1593462 1601594 5 1591592,1593443,1595383,1599191,1600658, 1591769,1594224,1595532,1599432,1602514, Q02040 uc004fpn.4

Surely this must be a mistake -- these genes can't be both on the X and Y chromosomes at the exact same coordinates, right? 

--Alex

[Luis Nassar]

Hello Alex,

Thank for your interest in the Genome Browser and for taking the time to write in.

You are correct that the knownCanonical table for hg38 contains some isoforms with duplicate gene symbols. This table is a subset of the GENCODEv29 data, which uses Ensembl IDs as primary keys (e.g. ENSG*). For hg38 knownCanonical, one canonical isoform per cluster ID means that there is only one isoform per Ensembl gene ID. The single isoform per gene ID is chosen by the parameters specified in the documentation (APRIS tag > GENCODE Basic set > longest isoform).

We do not have a direct way to extract a gene list with no gene symbol duplicates, that would be more difficult to produce and support due to gene symbol versioning and fluidity. If you would like to generate one, the knownCanonical table would be a good place to start, then filtering for the longest isoform is a common approach.

As far as the gene present on both the X and Y chromosomes, that example falls in a pseudoautosomal region (PAR1). These are regions that recombine during meiosis. The assemblers of the human genome deal with PARs by assembling a single haplotype sequence (as they do for all the autosomes), but then duplicating it in corresponding regions in X and Y. For some analyses such as sequence alignments, the PAR sequences are replaced with N's in one of the chromosomes so that the duplication of sequence doesn't mess up the results. The Genome Browser does not replace any sequence with N's. A list of PARs in hg38 can be found in the GRC website: https://www.ncbi.nlm.nih.gov/grc/human

I hope this is helpful. If you have any further questions, please reply to gen...@soe.ucsc.edu. All messages sent to that address are archived on a publicly-accessible Google Groups forum. If your question includes sensitive data, you may send it instead to genom...@soe.ucsc.edu.

Lou Nassar
UCSC Genomics Institute

Training videos & resources: http://genome.ucsc.edu/training/index.html
Want to share the Browser with colleagues?
Host a workshop: http://bit.ly/ucscTraining

--

---
You received this message because you are subscribed to the Google Groups "UCSC Genome Browser Public Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genome+un...@soe.ucsc.edu.
To view this discussion on the web visit https://groups.google.com/a/soe.ucsc.edu/d/msgid/genome/CAAOX%3DqG96DX%2BpB8%3D6HoGV_2yX0ANpRmUrbhxUQ6%2BY8hTpaRvKA%40mail.gmail.com.

[Alex Lenail]

Thanks for getting back to me, Lou. 

I've defaulted to choosing the longest of the canonical isoforms as the "true canonical" isoform for each gene symbol. 

For your records, I've attached the ~800 gene symbols for which the UCSC knownCanonical table lists multiple canonical isoforms (which are not from PAR regions). I've also copied the first few rows below -- the last column is boolean "canonical" field.

If there is another piece of biology which explains these (analogous to the PAR regions, which I hadn't known about before this, e.g. maybe something related to pseudogenes, etc...) I'd love to know!

	name	chrom	strand	txStart	txEnd	cdsStart	cdsEnd	exonCount	exonStarts	exonEnds	geneSymbol
0	ENST00000508969.2	chr11	-1	71810300	71813859	71810300	71810300	2	71810300,71813342,	71811948,71813859,	ALG1L9P	True
1	ENST00000532875.1	chr11	-1	71800540	71804640	71800540	71800540	4	71800540,71801779,71802857,71804529,	71800787,71801855,71802972,71804640,	ALG1L9P	True
2	ENST00000295900.10	chr3	1	63864559	64003462	63912598	63999467	13	63864559,63898398,63912587,63913156,63952378,6...	63864999,63898497,63912923,63913225,63952483,6...	ATXN7	True
3	ENST00000487717.5	chr3	1	63911928	64000207	63912598	63999467	12	63911928,63912587,63913156,63952378,63979914,6...	63912181,63912923,63913225,63952483,63980167,6...	ATXN7	True
4	ENST00000453174.7	chr10	1	79904897	79951029	79904897	79904897	8	79904897,79906571,79907667,79920951,79921913,7...	79905049,79906780,79907856,79921068,79921988,7...	BMS1P21	True
5	ENST00000634565.1	chr10	1	79906604	79907856	79906604	79906604	2	79906604,79907667,	79906780,79907856,	BMS1P21	True
6	ENST00000580790.1	chr10	-1	73699150	73730487	73699150	73699150	12	73699150,73704985,73713203,73718014,73719159,7...	73699588,73705078,73713332,73718165,73719299,7...	BMS1P4	True
7	ENST00000584747.5	chr10	-1	73715842	73730469	73715842	73715842	10	73715842,73718014,73719159,73720537,73721361,7...	73716180,73718165,73719299,73720659,73721504,7...	BMS1P4	True
8	ENST00000442201.6	chr3	-1	180614587	180679500	180614920	180679380	20	180614587,180616280,180616515,180616825,180619...	180615077,180616363,180616695,180616966,180619...	CCDC39	True
9	ENST00000476379.5	chr3	-1	180614007	180737874	180619347	180679380	25	180614007,180616280,180616515,180616825,180619...	180615077,180616363,180616695,180616966,180619...	CCDC39	True
10	ENST00000580501.2	chr10	-1	47581210	47582321	47581210	47581210	3	47581210,47581458,47582195,	47581358,47581578,47582321,	CTSLP2	True
11	ENST00000628708.1	chr10	1	46753604	46758198	46753604	46753604	3	46753604,46754833,46758077,	46753917,46754892,46758198,	CTSLP2	True
12	ENST00000425346.5	chr3	1	50350694	50354079	50351433	50353744	4	50350694,50351408,50352008,50353240,	50350984,50351560,50352046,50354079,	CYB561D2	True
13	ENST00000607121.5	chr3	1	50365373	50368197	50365373	50365373	3	50365373,50366296,50367795,	50365940,50366331,50368197,	CYB561D2	True
14	ENST00000443649.8	chr12	-1	122207662	122227534	122208380	122226014	7	122207662,122216487,122216758,122218265,122224...	122208577,122216584,122216869,122218397,122224...	DIABLO	True
15	ENST00000464942.7	chr12	-1	122207667	122226052	122208380	122226014	6	122207667,122216487,122216758,122218265,122224...	122208577,122216584,122216869,122218397,122224...	DIABLO	True

Happy 4th,

--Alex