How to estimate age of a whole genome seq?

[halffedelf]

hi I need to estimate the evolutionary age of a whole genome sequence. that is, I altered the reference human genome in certain aspects and trying to figure out the age of my altered genome sequence. I am not sure how I can do it, can't find any help in any literature. Should it be like the same way we do for estimating age of any gene, i.e., calculating the neutral mutations and multiply with million years per mutation? That case, which sequence should I compare with, the reference genome?

Any help/idea/suggestion/comment/feedback would by HIGHLY appreciated!

Thanks

[jme]

I don't do this, so disregard if this sounds silly, but surely if you're just modulating the reference (presumably to match an incomplete older genome), then wont comparing this new/modified version against the complete original reference give somewhat misleading results?

Depending on how many changes you've made, most of what you'll be comparing is same with same, and not indicative of what the genome you're trying to compare is(/was) like.

I imagine that conceptually you'd only need to compare the regions that you've altered to the same regions in the reference.

(Again, not my area so may be nonsense)

[Lluc]

What I understand by the "evolutionary age" of a gene is how long ago it appeared for the first time; and I would estimate it by looking at its phylogenetic distribution. What you are talking about is distances between homologous sequences; that is, time to their most recent common ancestor, right? If you think in terms of a tree, and you identify in the tree the time-point you want to estimate, you probably will know what sequence you need to use to compare yours with.

[halffedelf]

I don't do this, so disregard if this sounds silly, but surely if you're just modulating the reference (presumably to match an incomplete older genome), then wont comparing this new/modified version against the complete original reference give somewhat misleading results?

Depending on how many changes you've made, most of what you'll be comparing is same with same, and not indicative of what the genome you're trying to compare is(/was) like.

I imagine that conceptually you'd only need to compare the regions that you've altered to the same regions in the reference.

(Again, not my area so may be nonsense)

Thanks for your input. I understand what you mean. You are right in the sense that since I am altering the reference genome, comparing it with the reference genome will probably give me no result. I rather should probably consider comparing with chimp.

[halffedelf]

What I understand by the "evolutionary age" of a gene is how long ago it appeared for the first time; and I would estimate it by looking at its phylogenetic distribution. What you are talking about is distances between homologous sequences; that is, time to their most recent common ancestor, right? If you think in terms of a tree, and you identify in the tree the time-point you want to estimate, you probably will know what sequence you need to use to compare yours with.

Thanks for your input. I wonder if thats how people estimate the age from a whole genome sequence? For example, when they sequenced the whole genome of Neandertals, did they estimate the age by constructing a phylogenetic tree? I have estimated evolutionary ages of individual genomic components, e.g., an Alu or a gene, by counting the number of neutral SNPs among multiple individuals (that is essentially to find out the common ancestor of that genomic component), but when it comes to whole genome, I only have my altered whole genome, the reference genome and chimp. My understanding from your comment is that I can estimate the age of my altered whole genome by constructing a phylogenetic tree between these three, and using the knowledge that reference genome and chimp genome diverged 6mya. Thats a nice insight, I can definitely try doing that!

[halffedelf]

I wonder how did other groups do that, for example, when Paabo et al sequenced the neanderthal genome, how could they estimate the age of the genome? Either for constructing a phylogenetic tree between closely related genomes or counting the number of neutral SNPs, its a huge daunting task from computational point of view.

[Lluc]

What I meant is that it is not clear to me what you mean by "the age of a genome". Is it how long ago it diverged from a relative? Then, what relative? I don't think you need to build a tree. Just compare the genome you want to date with the relative from which you want to know how long ago it diverged.

Otherwise, if "the age of a genome" means "when the individual carrying this exact genome was alive", you'd better trust the estimated age of the biological sample from which the sequence was obtained, whether it is a Neanderthal's bone or a piece of salami.

If you read the paper and see how they dated the Neanderthal's genome, please, share it with us.

[halffedelf]

I meant "how long ago it was diverged". Thats where I am having difficulty, because if I had multiple polymorphic version of my genome, I could align them to find how long ago they diverged from the common anc, but I have only one related genome which is the human reference genome. The next closest fully assembled genome would be of chimp.

I found how Green et al estimated the age of Neandertals. Its a massive computational task. In summary, they aligned Neandertals, human ref and chimp ref with one another and identified the mutations specific to each species and went on from there. Instead of working on the whole genome, they had to remove the common regions before aligning which was another huge task. Its waaaay more complicated than I thought, but I am planning to give it a shot

If anyones interested in reading what Green et al did, the method is somewhat described in suppl text 10 of their paper. I am also open to discussion assuming it will help everyone else on related field.