Skip to content

Latest commit

 

History

History
113 lines (83 loc) · 4.89 KB

reference_material.md

File metadata and controls

113 lines (83 loc) · 4.89 KB

Guide to genomic reference material in Bento

This document includes a bit of information on what a reference genome is, what some common ones for Bento projects are, and some subtleties that are important to pay attention to.

Table of Contents

What is a reference genome?

Organisms of the same species share the vast majority of their DNA sequence, with variation in phenotype (appearance, behaviour, disease susceptibility, etc.) being to a large extent the result of variation in specific loci (regions) of the individuals' genome.

As a result, bioinformatics file formats often (but not always!) use a common set of sequences called a "reference genome", and data is associated with a particular chromosome and position within this genome.

As an example, one reference genome for humans is called "hg38"/"GRCh38". Human individuals usually have 46 chromosomes: 22 pairs of autosomes, and 1 pair of sex chromosomes (n = (22 + 1); 2n = 46 - see Figure 1). One copy of each chromosome is inherited from an individual's mother, the other from their father. The hg38 reference genome has 24 main genomic sequences: chr1-chr22 + chrX + chrY (Table 1); we don't need pairs here, since the pairs are extremely similar to one another; any variation between them can (usually) be well-represented

A karyotype of a human male

Figure 1: A karyotype of a human male with a normal chromosome configuration.

hg38 chromosome length
chr1 248956422
chr2 242193529
chr3 198295559
chr4 190214555
chr5 181538259
chr6 170805979
chr7 159345973
chr8 145138636
chr9 138394717
chr10 133797422
chr11 135086622
chr12 133275309
chr13 114364328
chr14 107043718
chr15 101991189
chr16 90338345
chr17 83257441
chr18 80373285
chr19 58617616
chr20 64444167
chr21 46709983
chr22 50818468
chrX 156040895
chrY 57227415

Table 1: hg38 reference genome chromosome lengths.

It's worth noting that the nuclear genome, i.e., what is described above, is not the only genome in human cells. Humans also have mitochrondria, which have their own circular genome. This genome, sometimes represented as chrMT in reference genome files, is 16569 bp long in humans.

File representation of reference genomes

Reference genomes are usually represented using the FASTA file format.

Our own Bento reference service ingests FASTA files, and does all operations using the FASTA format + an index file.

Common reference genomes

Here are some common reference genomes, and which Bento projects use them.

Human

  • CHM13v2.0: The "latest and greatest"; a more contiguous reference with several gaps filled in. This genome will be used in more projects going forward.
  • hg38/GRCh38:
    • BQC19: The CVMFS version on Alliance clusters at /cvmfs/soft.mugqic/CentOS6/genomes/species/Homo_sapiens.GRCh38/genome/Homo_sapiens.GRCh38.fa
  • hg19/GRCh37/hs37d5:
    • ICHANGE: The special hs37d5 version derived from GRCh37 is the 1000 Genomes Phase 2 reference genome. It is on Alliance clusters at /lustre03/project/6007512/C3G/analyste_dev/genomes/species/Homo_sapiens.hs37d5/genome/Homo_sapiens.hs37d5.fa.

Mouse

  • mm10/GRCm38: an older mouse (Mus musculus) reference genome.
    • ICHANGE: The CVMFS version on Alliance clusters at /cvmfs/soft.mugqic/CentOS6/genomes/species/Mus_musculus.GRCm38/genome/Mus_musculus.GRCm38.fa

Polar bear

Aspects to pay attention to!

  • There can be many different versions of the same base reference genome - for example, ICHANGE specifically uses the hs37d5 version of GRCh37, which DOES NOT have chr prefixes for the chromosomes, i.e., chromosome 1 is represented as 1 instead of chr1. It is important to check which specific version is used!

    VCFs can be queried to find evidence for which reference genome was used:

    bcftools view -h ./path/to/the.vcf.gz