Hi,
I downloaded the NCBI Refseq curated file of Genes and Gene Predictions from the UCSC Table Browser for hg38 as I want to use the exon coordinates as a target file for calling variants on Exome Sequencing data.
I noticed however, that the exon coordinates cover approximately double the genomic region as the exon coordinates in hg19 did (~80 million bps vs ~40 million). Is it possible that the size of the exome is really double in hg38?
I do not want to call variants on all of these regions since ~30% of these exonic regions are not covered at all in my WES data and another ~10% is covered by <10x. I would definitely like to exclude these regions from the target file but I do not fully understand what these regions are/why they were included in the first place.
Any help would be greatly appreciated.
•
written
by
ej
1 answer
I am not entirely sure why you are seeing such a huge difference in exome sizes between hg38 and hg19. Could you please describe in a little more detail how you are computing these values?
As far as RefSeq data are concerned, I strongly recommend you to download the relevant files from RefSeq and not UCSC. The data displayed in the UCSC browser are processed by the folks at UCSC and don't necessarily match RefSeq data exactly.
hg19 or GRCh37
RefSeq no longer actively annotates hg19 though updates are released occasionally. For the latest annotation data, go to NCBI Assembly and search for GRCh37. In the result page, click on the 'Download' button in the result card and choose RefSeq as source and GFF3 as your file format to download the latest version of annotation (released in September 2019) in GFF3 format.
hg38 or GRCh38
To download annotation for hg38 or GRCh38, go to NCBI Assembly and search for GRCh38. In the result page, click on the first hit to go to this page and use the blue 'Download Assemblies' button to download the RefSeq GFF3 file.
You will notice that RefSeq annotation data are provided in other file formats such as GTF and FASTA by following the same steps mentioned above.
Log in to answer this question.
-
target vs. bait BED files for Illumina WES
written by Juan Cordero
-
Tumor mutation burden calculation based on bed file?
written by lincaijin1994 •
-
How to transform a whole-genome callset into whole-exome callset?
written by Alejandro Rojo •
-
Trying to figure out coordinates for long range LD regions on build 38 for flashpca
written by curious
-
Using hg38 assembly with hg19 exome capture data
written by jdm204 •
-
Merge Samtools and Freebayes vcf files
written by kspata
-
How to get the no-overlapping exome region (.bed file)?
written by weiyuchung •
-
Somatic variants in exome normal/tumor data mostly in non-exonic regions. Is this a likely scenario?
written by ivivek_ngs
-
Annotation using annovar
written by ivivek_ngs
-
How To Understand Some Annotated Terms Made By Annovar
written by zengtony743
The default for hg19 is UCSC genes which is pretty much RefSeq and default to hg38 is GENCODE. GENCODE is more comprehensive and contains more genes, especially non-coding genes. For a fair comparison you have to compare the annotations from the same source for the two genome versions.
Thank you for your reply. In the UCSC Table Browser I chose to download the NCBI RefSeq set for hg38, not the GENCODE. Is this not from the same source as the NCBI RefSeq in hg19? Where can I find RefSeq coordinates for hg38?
No, exons should not vary that much from freeze to freeze. But, more importantly, if this is really about exome coverage then use the bed file that came with your kit. If you are interested in coding variants then use CCDS.