15
$\begingroup$

I have heard that a reference genome such as humans is generated by randomly choosing samples from a group of donors. But why do we call the DNA sequence generated as a reference? Why should we believe those few samples can represent all humans, from which we need to align with?

$\endgroup$
  • 2
    $\begingroup$ Good question. It is a quirk of history - the initial genome sequencing work was WAY too expensive and WAY too laborious to even contemplate sequencing multiple genomes. Thus, an initial genome (later extended to a couple of genomes) was used as a "reference". However, technological advances mean we can sequence genomes for under $1000 in a day now. Thus, the advent of things such as the 1000 genomes project (en.wikipedia.org/wiki/1000_Genomes_Project). $\endgroup$ – Cantona's Collar Apr 5 '15 at 23:25
  • $\begingroup$ Having said that, as the answer below points out, the variation between humans is tiny compared to the overall similarities. Thus, a "reference" can be generated from a sample of one or few. $\endgroup$ – Cantona's Collar Apr 5 '15 at 23:28
13
$\begingroup$

The main reason is because the genetic differences between individuals of the same species are tiny. For the vast majority of studies, they can simply be ignored.

Differences between individuals are usually (not always, but mostly) differences in SNP genotypes. These are single nucleotide differences which, while they can have phenotypic effect, don't really affect the ability of sequences to align. Sequence alignment software is designed to be able to deal with such small differences. For example, consider these two random sequences:

>seq1
ACCTTGCATCGGATCGAATTCGCGTTAGCGATCG
>seq2
GCCTAGCATCGGACCGAATTCCCGTTAGCAATCG

If we align them, we will get:

seq1            ACCTTGCATCGGATCGAATTCGCGTTAGCGATCG
seq2            GCCTAGCATCGGACCGAATTCCCGTTAGCAATCG
                 *** ******** ******* ******* ****

As you can see, despite the small differences in sequence, the two can be aligned very well. The same is true when you look at real data. My haemoglobin gene, for example, will align perfectly well against both yours and the one in the reference genome. In fact, they will almost certainly be identical or, at best, have a couple of different residues.

So, when doing work that is not about studying specific polymorphisms, we need a representative genome. Anyone's will do, unless they have a particularly serious mutation such as a chromosomal translocation.

$\endgroup$
  • 3
    $\begingroup$ To put some number, the average homology between human genomes is approx. 99.5%. As @terdon described, this relatively high homology (i.e. ~2,850,000,000 base pairs out of the total 3,000,000,000bp of the human genome will be identical) allows to align any human sequence to this so-called reference. This is of course in the scenario of no major chromosomal translocation, which usually leads to diseases. $\endgroup$ – cagliari2005 Apr 4 '15 at 15:44
  • $\begingroup$ @vajra78 it probably never will actually. These differences are irrelevant for most studies. You almost never care about individual genotypes, not unless this is a medical issue involving a specific individual. The point here is alignment and that is not affected by such small changes. Studies at the species level (looking at evolution or genetic regulation etc) rarely, if ever, need to take into account the intra-species variation. $\endgroup$ – terdon Apr 7 '15 at 10:37
0
$\begingroup$

Reference genomes do not accurately represent the set of genes of any single person.it is created by fragments of various donors, which when built,is used as a template for creating the real genome. Though we will find All humans are 99.9 per cent identical and, of that tiny 0.1 per cent difference the reference genome might not be 100% identical to the real genome.Every reference genome con be sequenced again and the gaps can be reduced each time. We can reach a point where we can say that the accuracy is high enough to be used as a reference platform for constructing the real genome

$\endgroup$
0
$\begingroup$

The same question sticked in my mind a month ago. Even @terdon gave an explanatory answer, I want to make a small contribution.

When I asked similar question on seqanswers *, one of users gave the link of the study Ashley Lab in Stanford. They generated the "Synthetic major allele human reference genomes". (1)

They combined the current reference sequence (in that time it was hg19), with the 1000Genomes data of variants with high allele frequencies. This combination creates three different reference genome of three human population (YRI, CEU and CHBJPT).

This synthetic reference sequences represents the variants that are highly seen in these population. It is possible to download these genomes from their page.

Dewey FE, et. al. (2011) Phased whole-genome genetic risk in a family quartet using a major allele reference sequence. PLoS Genetics 7(9): e1002280. http://dx.doi.org/10.1371/journal.pgen.1002280

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.