Hacker News new | past | comments | ask | show | jobs | submit login
An ancestral recombination graph of human, Neanderthal, and Denisovan genomes (advances.sciencemag.org)
49 points by diodorus 4 days ago | hide | past | favorite | 4 comments

As someone who knows next to nothing about genetics, this sentence confuses me a bit:

> We find that only 1.5 to 7% of the modern human genome is uniquely human.

I've read many times that "humans and orangutans share approximately 97% of their DNA" and other similar claims [1]. Knowing that, what makes this revelation so surprising?

[1]: https://www.nih.gov/news-events/nih-research-matters/orangut...

Repeated corroboration and validation is a critical part of science. They could have framed it differently such as "we reaffirm prev findings with our new method". This is good because it both validates previous findings and the new method. More precisely they are mutual consistent, which doesn't fully exclude the chance that all methods are biased in the same wrong way. However, the more independent validations there are, this alternative explanation is diminished in likelihood.

The statistic you are referring to when people quote % differences is on a protein level (in other words, genes -> proteins).Its not a very good metric and not used any more because it's very hard to quantify differences. See this article https://bmcgenomics.biomedcentral.com/articles/10.1186/s1286...

Our evolutionary history is complicated because we mated with many hominins and to make things even more complicated there different degrees of mixing throughout different geographical regions.

So as we (homo sapeins) evolved we mixed with these other closely related species. Chunks of our dna mixed with other hominins.

> We find that approximately 7% of the human autosomal genome is human-unique and free of both admixture and ILS. Roughly 50% of the human genome contains regions where one or more humans has archaic ancestry obtained through admixture.

They go on to talk about ~7% of haplotype blocks are uniquely human. The others have signatures of admixing. It doesn't mean the whole block is non-human but only that it has signatures of admixing.

Let's just say, genomicists really do play fast and loose with these sorts of things. For exmaple, with the human/ape comparison, they only compared highly conserved regions and totally ignored regions that didn't match up. So that's just totally misleading because it's heavily biased.

This work is just attempting to use probabilistic reconstruction to determine the tree topology of genetic history, the pull quotes are sort of to make it sound more accessible to the public. The underlying reality is that this is an ongoing area of research with a lot of contention and most of the ideas are actually undersupported by the evidence, but scientists gotta publish.

(I worked with one of the coauthors when he was a grad student)

Applications are open for YC Winter 2022

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact