(see the paper http://www.nature.com/nature/journal/vaop/ncurrent/full/nature12788.html
and Nature news on the surprise here: http://www.nature.com/news/hominin-dna-baffles-experts-1.14294)
My talk last October at the CAS-MPG Partner Institute for Computational Biology in Shanghai.
|Seminar Talk by Prof. Shi Huang|
|Author：webmaster News time：2012-10-22|
Time: 13:00-14:45pm, Oct. 25, 2012
Place: Room 223, SIBS main building
Host: Dr. Shuhua Xu
Title: Evidence for the Multiregional hypothesis of modern human origins from phylogeny-informative sequences
Abstract:Recent studies support the Maximum Genetic Diversity hypothesis that nucleotide diversities in fast evolving sequences are mostly at optimum level and hence cannot be informative to phylogeny inference. Consistently, here we found that fast evolving sequences are enriched with coincident or overlapping substitutions in humans and chimpanzees and also with shared SNPs among human races. We therefore analyzed the phylogeny-informative slow evolving sequences to revisit the question of modern human origins. We found that Europeans have greater nucleotide diversity than East Asians and Africans. The split time between Europeans and non-Europeans and between East Asians and Africans were estimated to be ~1.98 and ~1.78 Myr ago, respectively. East Asians are slightly closer to Europeans than Africans are, confirming genetic exchange between East Asians and Europeans as previously revealed by morphological studies. Europeans show distinct SNP patterns in non-synonymous SNPs from non-Europeans. Denisovans may be Africa H. heidelbergensis who may have migrated to Europe ~0.38 Myr ago and interbred with female H. antecessor in Europe to give rise to Neanderthals who had predominantly Denisovan nuclear genomes. These molecular results are consistent with fossil records and the Multiregional hypothesis.
This October I gave a talk on human origins at the HGV2013 in Seoul, Korea. http://www.hgv2013.org/program/schedule.html Here below is the abstract. (Ironically and a surprise to us, the same abstract was rejected even a chance for a poster presentation at the ASHG2013 meeting in Boston. But we did present two other abstracts, one on Autism and the other on transgenerational inheritance in yeast).
Understanding human evolutionary history may help solve health problems today. Using the 1000 genomes and other public sequences, we show that genetic diversity patterns of various human groups in most types of SNPs, including nonsyn sites in fast evolving proteins, syn sites in any proteins, non-coding, and mtDNAs, are all broadly similar to that of functional SNPs such as those altering/creating stop codons and splicing sites, thereby indicating purifying selection for most common SNPs and rendering them non-informative for phylogenetics. No population has uniformly lowest diversity in all types of SNPs, with East Asians lowest in non-coding, fast non-syn, slow non-syn, and syn sites, Europeans in stop codon, splicing sites, and mtDNA, and YRI in Y chromosome. Furthermore, shared SNPs among different populations mostly cluster in fast evolving DNAs consistent with independent mutations repeatedly hitting the same sites, invalidating the infinite sites assumption. Only slow evolving nonsyn SNPs that are theoretically predicted to be phylogenetically informative show the expected higher diversity for admixed populations relative to populations that contributed to such admixture, e.g., African Americans relative to Africans and Europeans. These 4923 SNPs are located in 178 proteins with much slower evolutionary rates than the average of those known to be under positive selection. We estimated a very conservative date for the split of Africans and non-Africans to be no less than one million years ago, which could be easily extended by 2-fold pending better determination of the human mutation rate in these slow evolving sites. Our analysis also provides data for Denisovans and Neanderthals as descendants of interbreeding between San and Iberians and as ancestors of Papuan and Australian aborigines, thus genetically linking all speakers of the click language. This molecular study presents a unified view of the complex multiregional evolutionary history of modern humans.
Also, the following abstract has been accepted for presentation (do not know if talk or poster yet) at the Cell Symposium on Human Origins in March 2014, Spain. Given our surprisingly negative experience at ASHG2013, we must give special thanks to the organizers of the Cell Symposium for their open mindedness.
Prior work has indicated that nucleotide diversities are mostly at optimum levels. From this perspective, we studied the origins of humans using the 1000 genomes project and other public data. We found empirical evidence of purifying selection for most common SNPs but neutrality for slow evolving nonsyn SNPs. Such slow SNPs allowed us to estimate a conservative time of no less than 2.11 million years for the split of Europeans and non-Europeans. Denisovans and Neanderthals/Altai were descendants of interbreeding between Africans and Europeans and ancestors of Australian aborigines. Analysis of mtDNA and Y chr provided independent evidence for the autosomes results. We also found a farmers-associated SNP and what may be gene conversions in mtDNA in admixed humans. These results suggest a unified human origins model consistent with physical/cultural traits and fossil records, which explains the wide diversity of Australian aborigines.
In October this year in Beijing, researchers from China held a summit on nearly all relevant topics of anthropology. I gave a talk right after professor Li Jin of Fudan University, the chief proponent in China of the Out of Africa model. Professor/Academician Xinzhi Wu of Chinese Academy of Sciences (CAS), the chief proponent in China of the Multiregional model gave a keynote speech at the meeting. Below is a picture taken at the meeting. The impressively healthy 84 year old Xinzhi Wu is at the middle of the front row. To his left is Li Jin, Xing Gao of CAS, and me. A book in Chinese from talks at this meeting is coming out in Spring 2014.
Our paper on human origins should be ready for submission soon. But every time we were getting ready, some fossil DNA or other data would turn up in the literature, such as the 24000 year old Siberian genome and now this 400000 year old Heidelbergensis, which just kept adding to the large amount of evidence for our story. We just have to keep updating our manuscript, which is of course very pleasant to us. It is truly an exciting time to study human evolution thanks to the remarkable fossil sequencing effort of professor Paabo and others.
Just a reminder, we held our story or any story of human origins to the highest standard of science: no single contradiction allowed, period. No future findings of any kind should be a surprise to us. If it is, we would concede we had the wrong story. We could be this confident because ever since the main theme of our story emerged about 3 years ago from the analysis of low quality autosomes HapMap data, we have analyzed numerous other datasets and we have not been surprised once. The thing that does surprise us maybe just a bit is how lawful and coherent an evolution story could turn out to be when studied in the right way.