This is a basic question but I couldn't find an answer through a web search; hopefully this is the right place to ask. Is the number of base pairs in a particular chromosome the same in all individuals? For example if I take an X-chromosome from two random humans would I count exactly 155,270,560 base pairs in both cases? or are there mutations that would make one longer than the other? If they're not exactly the same, what's the range in length variation?

  • 8
    $\begingroup$ Not only is it not the same in two different individuals. It's not the same in one individual (over their lifespan or two different cells at any given point). en.wikipedia.org/wiki/Telomere Also see en.wikipedia.org/wiki/Insertion_(genetics) and en.wikipedia.org/wiki/Deletion_(genetics). $\endgroup$ – Adam Phelps Aug 9 '15 at 6:11
  • 1
    $\begingroup$ @Adam make an answer out of that $\endgroup$ – MattDMo Aug 9 '15 at 6:18
  • $\begingroup$ If anyone would like to use my comment to formulate an answer (or improve on the existing answer please feel free). I don't really have the time right now to put together a well referenced answer I would be happy with. $\endgroup$ – Adam Phelps Aug 9 '15 at 6:29
  • $\begingroup$ @ Adam, I will add it. $\endgroup$ – Galen Aug 9 '15 at 6:34

Welcome to Biology.SE.

if I take an X-chromosome from two random humans would I count exactly 155,270,560 base pairs in both cases

No, you would probably not find the exact same number of base pairs because mutations do no only change one nucleotide to another (what we call a substitution) but sometimes add or delete few (or sometimes many) nucleotides.

note, btw that you don't need to take two different individuals, you can just consider the two X chromosomes of a female (or any other pair of chromsom in any gender) and find this difference in the length of chromosomes.

what's the range in length variation?

Good question (+1)!

Telomere issue

Before I start, I want to make clear that I consider the length of those chromosomes at the moment of conception. Chromosomes will vary in length during the lifetime due to telomere reduction. I will not consider this in the following calculations. Also, some mutations directly introduce (or delete) a large number of nucleotides (transposable elements for example), I am not considering those mutations here, assuming they are rare in comparison to to single insertions and single deletions (this assumption might not hold!). So please really take the following with a grain of salt.

Let's make some messy calculations

In classical theoretical population genetics, we tend to consider mostly substitutions. But I can maybe try to make some extrapolation out of this work if you allow me to make some strong assumptions, use poor estimates of actual true values and using some non-rigorous mathematics! This is going to be ugly and not extremely trustful $\ddot \smile $.

Not explaining why this is true (it is a result coming from Coalescent Theory), the expected number of pairwise differences between two neutral sequences for a diploid population is $E[\pi] = 4\cdot N\cdot \mu$ (quite an impressively simple result), where $N$ is the population size (assuming panmictic population) and $\mu$ is the mutation rate for the whole sequence. Assuming a constant per site mutation rate of $\mu_s = 10^{-9}$. Knowing the length of the sequence of interest (chromosom X) $L ≈ 1.55 \cdot 10^8 $, the mutation rate for the whole sequence is $\mu = L\cdot \mu_s = 0.155$. Let's consider a population size of $N = 5 \cdot 10^7$ (the equations assume a panmictic population so I just took some value that felt more or less reasonable to me much smaller than the actual worldwide population size). Therefore the total number of substitutions should be $E[\pi] = 4 \cdot 5 \cdot 10^7 \cdot 0.155 ≈ 10^7$.

Now, let's assume that only a fraction of $\frac{1}{100}$ of the mutations bring variation in the number of nucleotides, we might want consider the value $\frac{1}{100} \cdot 10^7 = 10^3$. And because a mutation that deletes a nucleotide from a long sequence will rather diminish the number of variant in sequence length that increasing it, let's say that will divide this number by 10!... so I'd say that two typical X chromosome would differ in length by about 100 nucleotides.

I am sure that with some work one can come up with more rigorous calculations and a more accurate expectation. Intuitively, the result of 100 nucleotides doesn't sound totally crazy (it is not 1 nor 10^6 at least).

Also, one could probably use available sequence data to estimate this value.

  • 1
    $\begingroup$ In some sense yes but it is kinda misleading the way you put it. Mutations can bring variations in the length of chromosomes. Then, within the lifetime of an individual chromosome length varies a lot becuase at every mitosis, chromsomes can shorten (you'll need to read about chromosoomal replicate, telomeres, DNA strands direction, Okasaki fragments to understand why). So one such mechanisms explain variation as a function of the age of the two individual you sample while the other mechanism in independent of age and explain variation in the length of chromosomes that are present in the egg. $\endgroup$ – Remi.b Aug 9 '15 at 18:42
  • $\begingroup$ Thanks for the explanation. So it seems there are (at least) two different mechanisms that can change the length of a chromosome : (1) telomere reduction and (2) random base pair deletion (is addition also possible?). From what I read so far the first involves removing parts of a repetitive sequence of bases at the ends of the chromosomes. This seems like a more deterministic process than the second. I think the calculations you made account for the second type. If correct (100 out 155 million) the range is pretty small. $\endgroup$ – unknown Aug 9 '15 at 18:55
  • 1
    $\begingroup$ I calculated the expected difference in length between 2 chromsomes sampled at random from eggs (without considering telomere shrinkage during lifetime, which would force us to consider the distribution of age, the distribution of the number of mitosis given age and the expected telomere shrinkage at each mitosis). The expected total range of length in a given population is necessarily greater that the expected pairwise difference in length....maybe 10 or 100 times as much I'd say intuitively. $\endgroup$ – Remi.b Aug 9 '15 at 19:35
  • 1
    $\begingroup$ There are also transposable elements that can change loci in a genome. In various microbes, they can take up or remove plasmids. Retroviruses can incorporate their DNA into their hosts. $\endgroup$ – Galen Aug 9 '15 at 19:38
  • 1
    $\begingroup$ You should open a new post for your new question and check an answer if your question in this post has been answered (or just don't check anything if your question hasn't been answered yet). I am not quite sure what exactly is your question but if you haven't yet, you might want to first follow a course of bioinformatics. $\endgroup$ – Remi.b Aug 9 '15 at 20:55

Humans generally have similar nucleotide counts, but they are not often going to be exactly the same. What allows for this variation is multiply determined, but here are some ideas.

The neutralist hypothesis states that most genetic change is not subject to selection [1], and this is especially evident in humans where there is a lot of junk DNA outside of operons as well as inside introns. If you look up 'the' human genome sequence, you'll find that it is a consensus sequence of multiple individuals [2].

Almost 70 % of all possible substitutions at the third codon position are synonymous [1].

Although to-date I have not worked with human sequences, I find that the same gene, say 16S or cpn60, in plants, fungi, and bacteria can be different across species, and even individuals.

Finally, as Adam pointed out in the comments, our chromosomes have regions of highly repetitive nucleotide sequence [3]. According to one source [4] on the telomere wiki page, vertebrates commonly have a repeating sequence of TTAGGG. Telomeres get shorter in the process of cellular mitosis, and this has been hypothesized to be from telomeric nucleotides being used as raw material for error correction in DNA replication.

[1] Bioinformatics: Genes, Proteins, and Computers. C.A. Orengo, D.T. Jones & J.M. Thorton

[2] http://www.sciencemag.org/content/291/5507/1304.full

[3] https://en.wikipedia.org/wiki/Telomere

[4] Sadava, D., Hillis, D., Heller, C., & Berenbaum, M. (2011). Life: The science of biology. (9th ed.) Sunderland, MA: Sinauer Associates Inc.

  • 1
    $\begingroup$ Thanks for the answer. Interesting comment about the error correction hypothesis; (I have some communication theory background where error correction is critical). As a side, when I was reading the description of the tolemere I was thinking it looked a lot a "preamble" whatis.techtarget.com/definition/preamble $\endgroup$ – unknown Aug 9 '15 at 19:13
  • $\begingroup$ What process would you speculate the telomere be a preamble for? In DNA replication, the origins of replication are not in the telomeres, maybe because the changes that occur in telomere sequences would silence the initiation sequences for DNA polymerase. An aside, genes sometimes become pseudogenes when mutations occur in initiation sequences of RNA polymerase. $\endgroup$ – Galen Aug 9 '15 at 19:47
  • 2
    $\begingroup$ These are the features of a preamble that brought out the comment : (1)Preamble doesn't contain any information; its purpose is to tell you when to start looking for information. It is also used to tell you if there's something there are at all. (2)Preamble is usually very simple and repetitive. (3)Its exact length doesn't matter but the shorter it gets, the less useful it gets.In noisy channels, a longer preamble is easier to detect and process. (BTW, I really don't have any hypothesis here; for all I know these are just accidental similarities). $\endgroup$ – unknown Aug 9 '15 at 21:02

Are there mutations that would make one longer than the other? If they're not exactly the same, what's the range in length variation?

As pointed out in the other answers, there can be mutations that change the length of the DNA between individuals and even single cells of the same tissue.

There are different processes that can change the genome length. The most common ones are:

We can ignore microindels and chromosomal translocations for the moment because the former impart only a small change to the genome size and the latter are usually rare and in most cases, are extremely deleterious.

With a stringent mapping, CNVs are known to cause 4.8% variation in the size of the human genome [2]. It is to be noted that this information does not account for the frequency of the changes; this is just a sort of maximum limit. It can also be noted that deletions have a greater contribution to variation compared to insertions.

LINE-1 retrotransposons are known to be quite active in the human genome. They accounts for ~17% of the human genome and are one of the major sources of inter-individual variation [3]. They are especially noted for promoting somatic heterogeneity in differentiating neurons [4]. Since LINEs expand by copy-pasting, their activity is correlated with increase in genome size. Transposon activity is generally suppressed in germ cells by a class of small non-coding RNAs called piRNAs [5].

I haven't found an article that exactly describes the variance in the genome size due to LINE-1 or other transposable elements. I'll add a reference if I find one.

  • $\begingroup$ Interesting numbers. If one mechanism alone (CNVs) can cause 4.8% variation in length and some claim "in terms of DNA sequence all humans are 99.5% similar to any other humans" (en.wikipedia.org/wiki/Human_genetic_variation) then either some very loose definitions are being used or these variations have odd statistics – $\endgroup$ – unknown Aug 10 '15 at 21:28
  • 1
    $\begingroup$ @secretlyfamous Note that CNVs need not disrupt homology i.e. "similarity" between two genomes. When we talk about similarity we generally refer to sequence homology. And the statistics presented are for the upper-bound of the length of CNV regions. The variation is therefore an upper-bound and not the standard deviation. $\endgroup$ – WYSIWYG Aug 11 '15 at 4:53
  • $\begingroup$ thanks for the explanation. It's a little hard for me to accept that (unrestricted) insertion/deletion does not change homology; but I'll save the details for a future question. $\endgroup$ – unknown Aug 11 '15 at 6:03
  • 1
    $\begingroup$ @secretlyfamous imagine duplications, it just creates an extra copy of the genomic region which would again be homologous to the reference. There is a chance of homozygous deletions too. Another point to be noted is that when people say that the genome is 99% similar (or even when comparison between different species is made; for e.g. chimpanzee and human genomes are 95%(?) similar) then the non-transcribed (intergenic) regions are not considered. $\endgroup$ – WYSIWYG Aug 11 '15 at 6:10

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.