Reference genomes

Reference Genomes

• Fasta format

• Assembly version, generally, associated to a number and a date of assembly

• A same assembly may be provided by various organisation (Genome Resource Consortium, Ensembl, NCBI, UCSC, etc)

  This will be the same DNA sequence but formats may differ:

  • by the name of the chromosomes (chr1, 1, NC_000001.11, ...)
  • by the presence (or the absence) of unmapped contigs and haplotypes

exemple 1: human genome

• GRCh37/hg19 - juil 2007
• GRCh38/hg38 - déc 2011
• GRCh39/hg39 - juin 2020 (repeat ++)

This various versions (or "releases") may in addition contain

• chromosomal regions "Aplotypes" (HLA, HBV inserts, etc…)
• unmapped contigs (regions which are significant assembly of reads, but are not assigned to a specific chromosome)

exemple 2: mouse genome

Release name	Date of release	Equivalent UCSC version
GRCm39	June 2020	mm39
GRCm38	Dec 2011	mm10
NCBI Build 37	Jul 2007	mm9
NCBI Build 36	Feb 2006	mm8
NCBI Build 35	Aug 2005	mm7
NCBI Build 34	Mar 2005	mm6

Annotations

It is important to note that annotations of genomes (GTF, GFF, etc.) although generally equivalent, are strictly linked to their genome version because they refere to the DNA sequences using the format of the release. This is why a GTF annotation file downloaded from Ensembl is not interchangeable with a GTF annotation file from the UCSC or from another organisation.

Moreover, since genome annotations may be considered as genome metadata (data on data), it is normal and expected that genome annotation versions are different from genome versions and that they are released at a faster pace.