Archaic hominin traits through the splicing lens

Jun 05, 2023

Human evolution

Nature Ecology & Evolution volume 7, pages 800–801 (2023)Cite this article

2067 Accesses

15 Altmetric

Metrics details

Machine-learning-based prediction of splicing in extinct hominin species highlights the effect of natural selection on splice-altering variants and reveals phenotypic differences with modern humans.

The sequencing of the genomes of archaic hominins has fostered a renewed interest in the identity of our extinct relatives and their legacy in the genome of modern humans. However, characterization of the phenotypes of archaic hominins is limited by the scarcity of remains and the rapid degradation of soft tissues after death. Various attempts have been made to infer archaic phenotypes on the basis of their DNA methylation patterns1 or through the study of the regulatory alleles that they share with modern humans2. Yet, the extensive purge of archaic DNA from the genome of modern humans3 and the possibility of regulatory divergence between archaic and modern humans4 limit the effectiveness of these approaches. Writing in Nature Ecology & Evolution, Brand et al.5 implemented a solution to this issue by focusing on alternative splicing, which informs us on the phenotypes of archaic hominins and reveals how splice-altering variants (SAVs) that modern humans inherited from their archaic relatives helped them to adapt to their environment, but also contributed to disease (Fig. 1).

a, Brand et al. infer genetic differences between modern humans and four lineages of archaic hominin. b, They evaluate the effect on splicing of thousands of archaic alleles and show an increased deleteriousness of SAVs that are specific to one or more archaic lineages relative to those that are shared with modern humans. c,d, They further overlap these SAVs with disease genes to reveal specific traits of archaic humans (c) and explore the effect of introgressed SAVs on human phenotypes and adaptation to environmental pressures (d). Created with BioRender.com.

Alternative splicing — the process by which exons of a gene are joined in different combinations to form alternative mRNA molecules — is a major contributor to the heritability of complex traits, on a par with variants that affect gene expression levels6. Because molecular mechanisms that underlie alternative splicing are deeply conserved across eukaryotes7, Brand et al. reasoned that the same rules should govern alternative splicing in modern humans and archaic hominins. They used a deep-learning algorithm known as SpliceAI8 to predict allele-specific effects on splicing at more than 1.5 million loci that differ between modern humans and four archaic hominins (three Neanderthals and one Denisovan). The authors identified thousands of SAVs present in archaic hominins, of which 37% are specific to these hominins (referred to as archaic-specific SAVs). The remainder included both ancient alleles that predate the split between modern humans and their archaic relatives and have survived in modern humans to this day (59%) (referred to as ancient SAVs), as well as alleles of archaic origin that were introgressed into modern humans through admixture (4%) (referred to as introgressed SAVs).

The authors showed that archaic alleles that are specific to a single lineage (for example, specific to the Neanderthals from Vindija Cave) are enriched for SAVs compared to archaic alleles that are shared across multiple lineages of archaic hominin. This result, which they attribute to the purging of SAVs that appeared early in archaic hominin evolution, is consistent with SAVs being more frequently found within Neanderthal lineages. Indeed, Neanderthals had a smaller effective population size relative to Denisovans and natural selection was thus less efficient at purging deleterious alleles from their genomes. Brand et al. further showed that archaic-specific SAVs are more likely than ancient SAVs to alter conserved sites and disrupt protein function. Finally, the authors analysed the phenotypic effect of the genetic burden carried by our late archaic relatives, by intersecting archaic-specific SAVs with known disease genes (identified by genome-wide association studies or associated with rare Mendelian disorders). In doing so, they pinpointed the specific phenotypes that are associated with SAV-containing genes in each archaic lineage, such as fragile skin in Neanderthals or muscular abnormalities in Denisovans.

Next, Brand et al. focused on SAVs that were introgressed in modern humans. The authors found that introgressed SAVs are enriched near genes that are expressed in a tissue-specific manner and show that the vast majority of SAVs observed in humans today are older than the split between Denisovans and Neanderthals. These results suggest that natural selection purged more-recent SAVs from human genomes shortly after their introgression, leaving only SAVs with more-localized effects on gene expression. Yet, the authors further showed that introgressed SAVs are enriched near susceptibility genes for several phenotypes, including hay fever and allergies, Helicobacter pylori serological status or systemic sclerosis, which suggests that SAVs inherited from archaic hominins still contribute to disease today.

Finally, the authors explored how SAVs that modern humans inherited from their archaic relatives helped them to adapt to their environment. In addition to previously described SAVs associated with COVID-19 susceptibility at the OAS1 locus and rhinitis at the TLR1 locus, the authors report SAVs at the EPAS1 locus, where an Denisovan-introgressed haplotype has contributed to the adaptation of Tibetan peoples to high altitudes9. The reported SAV leads to nonsense-mediated decay of EPAS1 and correlates with lower haemoglobin levels in Tibetan individuals10, consistent with the positive effect of EPAS1 on altitude-induced erythropoiesis11. At the ERAP2 locus, the authors identify three introgressed SAVs of potential evolutionary importance. First, they identify an ancient SAV evolving under balancing selection that is associated with stimulation-induced splicing of ERAP212 and increased survival during the Black Death13. In addition, they report two Neanderthal-specific SAVs of unknown functional relevance, one of which is predicted to induce nonsense-mediated decay of ERAP2. Together, these results highlight how SAVs inherited from archaic hominins have contributed to human adaptation.

The work presented by Brand et al. shows us how the prediction of molecular phenotypes from DNA sequence8 can teach us about extinct species and their effect on the evolutionary history of their living relatives. Yet, it also has some limitations. First, it relies heavily on the accuracy of the underlying predictions. Its conclusions could thus suffer from the poor performance of current prediction methods when applied to deep intronic variants and distal enhancers. Second, Brand et al. only consider variants with a widespread effect on splicing. Yet, context-specific splicing has been shown to have an essential role in the development of the brain and testis14 and may have had a role in the purging of archaic variants observed in these tissues4. Although the general splicing machinery is extremely conserved, such dynamic and tissue-specific splicing may be much more variable across species, which may complicate the portability of predictions across hominins. More-accurate models powered by massively parallel splicing assays15 and the mapping of splice quantitative trait loci6 are needed to embrace the complexity of such context-specific splicing. Finally, structural variants (for example, insertions and/or deletions) and variants located on sex chromosomes have so far been left aside, despite their relevance for human phenotypes and hybrid infertility. Future efforts must assess the role of these variants in the purge of archaic alleles from modern human genomes.

The work of Brand et al. is a powerful proof of concept that lays the foundations for future studies that aim to resurrect the alternative splicing landscape of extinct species.

Gokhman, D. et al. Cell 179, 180–192 (2019).

Colbran, L. L. et al. Nat. Ecol. Evol. 3, 1598–1606 (2019).

Article PubMed PubMed Central Google Scholar

Harris, K. & Nielsen, R. Genetics 203, 881–891 (2016).

McCoy, R. C., Wakefield, J. & Akey, J. M. Cell 168, 916–927 (2017).

Article CAS PubMed PubMed Central Google Scholar

Brand, C. M., Colbran, L. L. & Capra, J. A. Nat. Ecol. Evol. https://doi.org/10.1038/s41559-023-02053-5 (2023).

Li, Y. I. et al. Science 352, 600–604 (2016).

Article CAS PubMed PubMed Central Google Scholar

Collins, L. & Penny, D. Mol. Biol. Evol. 22, 1053–1066 (2005).

Article CAS PubMed Google Scholar

Jaganathan, K. et al. Cell 176, 535–548 (2019).

Article CAS PubMed Google Scholar

Huerta-Sánchez, E. et al. Nature 512, 194–197 (2014).

Article PubMed PubMed Central Google Scholar

Beall, C. M. et al. Proc. Natl Acad. Sci. USA 107, 11459–11464 (2010).

Article CAS PubMed PubMed Central Google Scholar

Liu, H. et al. Blood Cells Mol. Dis. 84, 102446 (2020).

Article CAS PubMed Google Scholar

Ye, C. J. et al. Genome Res. 28, 1812–1825 (2018).

Article CAS PubMed PubMed Central Google Scholar

Klunk, J. et al. Nature 611, 312–319 (2022).

Article CAS PubMed PubMed Central Google Scholar

Mazin, P. V., Khaitovich, P., Cardoso-Moreira, M. & Kaessmann, H. Nat. Genet. 53, 925–934 (2021).

Article CAS PubMed PubMed Central Google Scholar

Rong, S. et al. Preprint at https://doi.org/10.1101/2022.11.20.515225 (2022).

Download references

Institut Pasteur, Université Paris Cité, CNRS UMR2000, Human Evolutionary Genetics Unit, Paris, France

Maxime Rotival

You can also search for this author in PubMed Google Scholar

Correspondence to Maxime Rotival.

The authors declare no competing interests.

Reprints and Permissions

Rotival, M. Archaic hominin traits through the splicing lens. Nat Ecol Evol 7, 800–801 (2023). https://doi.org/10.1038/s41559-023-02045-5

Download citation

Published: 04 May 2023

Issue Date: June 2023

DOI: https://doi.org/10.1038/s41559-023-02045-5

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative