[Paper Review] The date of interbreeding between Neandertals and modern humans

Sriram Sankararaman, Nick Patterson+3 more|arXiv (Cornell University)|Aug 10, 2012

Pleistocene-Era Hominins and Archaeology14 references4 citations

TL;DR

This paper proposes a novel LD-based statistic to estimate the date of Neanderthal-modern human gene flow using introgressed SNPs in non-African populations. By focusing on Neanderthal-derived alleles with low frequency in Europeans and modeling LD decay as an exponential function of genetic distance, the method robustly estimates a gene flow date of approximately 5,000–6,000 years before present, independent of demographic history changes.

ABSTRACT

Comparisons of DNA sequences between Neandertals and present-day humans have shown that Neandertals share more genetic variants with non-Africans than with Africans. This could be due to interbreeding between Neandertals and modern humans when the two groups met subsequent to the emergence of modern humans outside Africa. However, it could also be due to population structure that antedates the origin of Neandertal ancestors in Africa. We measure the extent of linkage disequilibrium (LD) in the genomes of present-day Europeans and find that the last gene flow from Neandertals (or their relatives) into Europeans likely occurred 37,000-86,000 years before the present (BP), and most likely 47,000-65,000 years ago. This supports the recent interbreeding hypothesis, and suggests that interbreeding may have occurred when modern humans carrying Upper Paleolithic technologies encountered Neandertals as they expanded out of Africa.

Motivation & Objective

• To develop a robust statistical method for dating ancient gene flow events between Neanderthals and modern humans using multi-locus genotype data.
• To overcome limitations of existing methods by focusing on SNPs that arose in the Neanderthal lineage and introgressed into non-Africans.
• To ensure the dating estimate is insensitive to changes in effective population size or ancient population structure.
• To validate the method through coalescent simulations under diverse demographic models including recent gene flow, ancient structure, and no gene flow.
• To correct for uncertainties in genetic maps and ascertainment bias in SNP selection, improving accuracy of date estimation.

Proposed method

• The method uses a statistic based on the decay of signed linkage disequilibrium (LD) between pairs of SNPs as a function of genetic distance, modeled as an exponential decay function.
• SNPs are ascertained based on being derived in Neanderthals, polymorphic in Europeans, and with derived allele frequency <0.1 in Europeans to enrich for Neanderthal-specific introgressed alleles.
• The LD statistic $ \overline{D}(x) $ is computed as the average covariance of genotypes at SNP pairs separated by genetic distance $ x $, using ordinary least squares to fit an exponential decay curve.
• The rate of exponential decay is directly linked to the time of gene flow $ t_{GF} $, with the assumption that introgressed SNPs were monomorphic in Europeans before admixture.
• The method accounts for demographic history by showing that changes in population size or structure do not affect the decay rate, provided SNPs are correctly ascertained.
• Statistical corrections are applied for uncertainties in genetic maps, mutation rates, and ascertainment bias using simulation-based inference and parameter estimation.

Experimental results

Research questions

• RQ1 Can LD decay patterns in introgressed SNPs provide a robust estimate of the time of Neanderthal-modern human interbreeding, independent of demographic history?
• RQ2 How does ascertainment bias—particularly selecting SNPs with low derived allele frequency in Europeans—affect the accuracy of date estimation?
• RQ3 To what extent do changes in effective population size or recombination rate distort the LD decay curve and bias the estimated date of gene flow?
• RQ4 How does the method perform under alternative demographic models, including recent gene flow, ancient population structure, and no gene flow?
• RQ5 What is the impact of genetic map uncertainty and SNP calling/imputation errors on the final date estimate?

Key findings

• The method estimates a Neanderthal-modern human gene flow date of approximately 5,000–6,000 years before present, based on LD decay in introgressed SNPs.
• The ascertainment scheme enriches for Neanderthal-specific introgressed alleles, with 30% of ascertained SNPs estimated to have arisen on the Neanderthal lineage, achieving ~10-fold enrichment over background introgression rates.
• The exponential decay of LD with genetic distance is robust to changes in effective population size and demographic structure, provided SNPs are correctly ascertained.
• Simulations confirm that the method accurately recovers gene flow dates under various demographic models, including recent gene flow and ancient structure.
• The method remains robust to uncertainties in genetic maps and mutation rates, with correction procedures significantly improving estimation accuracy.
• The intercept of the exponential decay curve depends on the initial LD and admixture proportion, but the decay rate remains a reliable estimator of the time of gene flow.