A long, long time ago, in a land far, far away (2014 at the Hebrew University of Jerusalem, Israel actually!) a team of researchers banded together under the leadership of Liran Carmel and David Gokhman. Rather than inventing the wheel or devising a strategy to capture wooly mammoth, the team developed a way to reconstruct DNA methylation maps from ancient DNA to highlight epigenetic differences at genes essential for limb development between modern humans (if we can call ourselves that!) and our ancient brethren – the well-known Neanderthals and the lesser-known Denisovans, whose morphology remains unknown.
Fast-forward to the modern era (2019, to be exact) and the epigenetic fossil-hunting duo are back leading a new study that employed ancient DNA methylation analysis to reconstruct prehistoric anatomies. Fascinatingly, the team used DNA methylation mapping to infer gene expression patterns, which they first utilized to reconstruct the skeletal morphology of chimps and Neanderthals to compare against known anatomy. Then, they used their method to propose how the ancient Denisovans may have looked.
So how did the team put a face to that very ancient of names through DNA methylation analysis?
- The method compares DNA methylation patterns
in Denisovans (derived from bone and teeth) to those of modern humans,
Neanderthals, and chimpanzees
- Differentially methylated regions (DMRs) that appear in the distinct species can infer gene regulation, and hence phenotype
- To reconstruct morphology with selected DMRs,
the authors required that:
- DMRs reflect lineage-specific changes – by filtering out loci whose methylation levels differ across skeletal parts, age, health state, or sex and choosing DMRs with a >50% change in methylation
- Lineage-specific DMR methylation changes reflect changes to gene activity – by selecting DMRs close to transcriptional start sites and confirming that DNA methylation variation leads to gene downregulation
- Said changes in gene expression should link to phenotypic effects – by seeking DMR-associated genes with a known loss-of-function phenotype using the Human Phenotype Ontology (HPO) database
- The authors then removed phenotypes not described on a one-dimensional phenotypic scale (e.g., higher-lower/accelerated-delayed) and, depending on the number and activity status of DMRs associated with a specific gene (and hence phenotype), the authors created an anatomical estimate for each hominin compared to modern humans
- With this information, the authors reconstructed
skeletal profiles for Neanderthals and chimpanzees
- Their strategy allowed the robust prediction of traits that separate both Neanderthals and chimpanzees from modern humans
- Application of this DNA methylation-led methodology
to Denisovans allows for the construction of a putative morphological profile
- Analysis suggests that Denisovans shared an elongated face and a wide pelvis with Neanderthals
- However, predicted Denisovan-associated changes include increased dental arch and lateral cranial expansion, which match both the only morphologically informative Denisovan bone discovered to date and the Xuchang skull (suggested as Denisovan)
Overall, this prehistoric study suggests that ancient DNA methylation profiles can be used to reconstruct anatomical features and put a face to some of the ancient names that lie in the long evolutionary history of modern human.
For more on how this fascinating new epigenetic study peers far back into an era long, long ago, see Cell, September 2019.