All humans have the same sort of proteins in our bodies. Take haemoglobin for example.

Is the gene coding for haemoglobin in my body identical to everyone else's gene or is there slight variations in the nucleotide sequence?

Are there examples of proteins that are always completely conserved at the population level?


Humans have many variants

There is variation. The project I use to help understand this natural variation is gnomAD. Using VarMap and a slightly out of date gnomAD file, I counted 16007805 protein-coding variants across the human genome. This number will only go up over time.

Indeed, the 1000 Genome project found that on average each person has between 250-300 loss of function protein variants that are not found in their parents (The 1000 Genomes Project Consortium, 2010).

This is an important concept for human health. ClinVar is one of many projects that aims to catalogue and study when these variations lead to disease in a clinical context. Another is the 100,000 genome project by Genomics England which studies NHS patient data in cases of rare disease and cancer.

Haemoglobin has variants, including disease variants

At the time of writing, HBA1 (haemoglobin alpha subunit gene) has 183 gnomAD variants and 17 pathogenic variants in ClinVar (sourced from gnomAD). Again, both of these numbers are likely to increase because the data will cover more people.

Constraints on highly important proteins

But the underlying question is, I think, "are there some proteins that are so important, that life keeps them highly constrained" i.e any variation will lead to an invalid cell or a disease phenotype. gnomAD attempts to add "constrained" metrics to each protein record, and some are more constrained than others.

For example:

  • Haemoglobin scores a pLI of 0.01 (higher scores are more intolerant to variation, specifically loss of function variation).

  • p53 is a gatekeeper of the cell cycle, mutants of which are common in cancer cells. It has a pLI score of 0.53 which means it is very intolerant to variation compared to haemoglobin.

  • Ribosomal protein L5 has a pLI of 0.998 implying it can tolerate little if any variation. The ribosome is critical in protein production, hence altering it may cause a complete breakdown of cellular life.

Variation and Evolution

There is an almost philosophical difference between human variation and human evolution. Variation is a static snapshot of our protein sequence from individual to individual. Evolution in the sense of a Dayhoff Matrix requires looking back millions or billions of years by comparing similar protein sequences across many species.


It is highly unlikely that there exist any protein that is made from completely identical nucleotide sequences across the entire human population. There will certainly be regions within a gene that are highly conserved, but there is little evolutionary pressure to conserve an entire gene's nucleotide sequence across the population.

This is in part due to the fact that different codons can translate into the same amino acid. So, even if a protein is (phenotypically) identical across the human population, the genes making that protein can use different sequences to code for the same amino acids. This alone can allow for an incredibly large number of possible gene variations that can code for an identical protein.

Related resources: Neutral mutation (wikipedia) and the neutral theory of molecular evolution (wikipedia)

  • 2
    $\begingroup$ Haemoglobin has two distinct chains. It would be better to give separate numbers for each. $\endgroup$ – David Sep 9 at 22:04
  • 2
    $\begingroup$ I think there are a lot more than 17 variants of human Hb! See A Database of Human Hemoglobin Variants and Thalassemias $\endgroup$ – user1136 Sep 9 at 23:27
  • 1
    $\begingroup$ The fact that different codons can code for the same amino acid doesn't give anywhere near "infinite possibility"! Most AAs are only coded for by 2 different codons, and only one is coded for by 6. Given a fixed length of gene, that is a far cry from infinite variability! $\endgroup$ – terdon Sep 10 at 9:02
  • 3
    $\begingroup$ @terdon Hmmm... Even only considering two different codons for each amino acids for a small 100 amino-acids proteins, that gives 2^100 possibilities, a little in excess of 10^30... Not infinite, sure, but rather high ! $\endgroup$ – Vincent Fourmond Sep 10 at 13:30
  • 3
    $\begingroup$ Anticodon are a funny example for conservation, because tRNA genes actually have a multitude of copies, and while itʼs true that thereʼs considerable negative selection on their anticodon loop, variability certainly exists — even here. $\endgroup$ – Konrad Rudolph Sep 10 at 14:09

At the whole-gene level, there is likely no absolute conservation of any human protein-coding gene at the population level, though there might be complete conservation between individuals. Keep in mind that most human genes are on the order of tens of thousands of base pairs long, and only a portion of that length encodes functional motifs. There are, however, ultra-conserved elements on the order of hundreds of base pairs that are identical between humans, rats, and mice. Most of these elements are non-coding, and some are transcribed as functional ncRNAs.

  • 4
    $\begingroup$ "At the whole-gene level, there is no absolute conservation of any human protein-coding gene at the population level," Are you sure about this? I wouldn't be surprised if there were a specific, very short gene with at least its coding sequence 100% conserved. Or even one with everything including UTRs conserved (perhaps a single-exon gene with no introns). I also wouldn't be surprised if no such thing existed, but you are explicitly ruling it out without a reference. $\endgroup$ – terdon Sep 10 at 9:05
  • $\begingroup$ @terdon you're right -- I don't give a reference. I should have qualified that first sentence with "As far as I know, ..." or, as Gen Test said, "It is highly unlikely ... " $\endgroup$ – Dirigible Sep 10 at 15:13
  • 2
    $\begingroup$ @terdon I haven't come across a whole protein that is completely "constrained", but there are indeed heavily constrained regions in both coding and non-coding regions. $\endgroup$ – James Sep 10 at 16:24
  • $\begingroup$ @James neither have I. If I had to bet I'd say that there is no case of a protein coding gene whose coding sequence is 100% identical across all (healthy) individuals, but this is the human genome, so nothing can be ruled out! :) $\endgroup$ – terdon Sep 10 at 16:41
  • $\begingroup$ @terdon My delay in response was me awk|sort|uniq-ing my gnomAD file. I have 18922 uniq UniProt codes, so there are a lot less than the complete proteome! That is more likely down to bugs than biology though! I'll report back when I am more confident. $\endgroup$ – James Sep 10 at 17:12

No gene and no base pair is immune to mutation. It is the natural selection pressure that keeps some gene relatively constant. Neutral mutations are not subject to pressure, so everything changes. And then, there are "almost neutral" mutations that also get passed by.


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.