Choice of ref. Genomes

Question

[cetin.m]

The human reference genomes are just somebody's genome. What kind of errors and biases does this introduce?

Can you please direct me to relevant websites or papers?

Answer 1

[Kevin Blighe]

Yes, GRCh37 / hg19, for example, is mostly based on a single individual from Buffalo, New York state. Take a look at my answer here, and that of Emily: A: Alternate nucleotide is more frequent than reference nucleotide. OMG I'm dizzy.

Note that GRCh38 / hg38 was constructed based on the data produced by the 1000 Genomes consortium; so, it is more 'wholly representative'. This said, I believe that the best procedure is to have multiple reference genomes that are tailoured for each ethnic region of the World.

Kevin

[cetin.m]

Thanks!!

Where can I find more information about how GRCh 38 was assembled?

I did google but no find.

[Kevin Blighe]

Take a look at this great post by Giovanni: The New Grch38 Human Genome Browser Has Arrived!

Answer 2

[igor]

There is some info about the biological source of the human reference from the Genome Reference Consortium FAQs:

The human reference genome is a composite genome, derived from the sequence of several different anonymous individuals. Approximately 93% of the GRCh38 primary assembly (the assembled chromosomes, unlocalized and unplaced sequences) consists of sequences from 11 genomic clone libraries (a library can generally be considered a proxy for an individual’s genome). One of these libraries, RP11 or RPCI - 11 Human Male BAC Library has a much higher representation than all others, and contributes to 70% of the primary assembly. The donor of RP11 library was an anonymous male, though analysis suggests his DNA is an African-European admixture (see page 146 of Supporting Online Material of PMID:20448178). The remaining 7% represents sequences from >50 libraries. These libraries were developed from individuals (male and female), as well as flow-sorted chromosomes from various cell lines. The make-up of GRCh37 is largely similar to GRCh38 (Figure 1).

There are many additional facts about the reference sequence on that page.