RuNIE Long-Form Bayesian MCMC Skyline & the 5-Stage Neanderthal Population Crash (650 to 39.2 ka BP): Coupled 95% Credible Chronogram, Pristine Phylogram, and Effective Population Size Trajectory Benchmarked Against *BEAST across an 87,227-bp Alignme

Coupled Metropolis-Hastings MCMC sampling (runie-mcmc-skyline: G = 10,000 generations, 20% burn-in = 2,000 discarded, 8,000 posterior states) across 64 ancient/modern hominin mitogenomes and 16 nuclear loci (L = 87,227 *MAFFT / 86,612 Native Rust gotoh-msa aligned bp) resolving the 5-stage Neanderthal Ne(t) collapse (13,893 to 602.7 at 39.2 ka BP) and 37.3x speedup over *MAFFT + *IQ-TREE + *BEAST v1.10.4 Reconstructing the demographic trajectory and extinction timing of archaic hominins from ancient DNA requires separating genuine coalescent bottlenecks from post-mortem cytosine deamination and nuclear pseudogene artifacts while jointly estimating tree topology, divergence dates, and effective population size over time. Here we introduce and validate a configurable long-form Bayesian Markov chain Monte Carlo sampler inside the Rust Native Inference Engine (runie-mcmc-skyline; default 10,000 generations with a 20% burn-in gate) and benchmark it against *BEAST, *IQ-TREE with *LSD2, and *PAML codeml across an 87,227-bp hominin alignment spanning Neanderthals, Denisovans, Sima de los Huesos, and modern humans. Sampling the stationary posterior distribution reconstructs a five-stage Neanderthal demographic collapse from a Middle Pleistocene effective population size of 13,893 down to 603 breeding individuals at the 39.2 ka extinction horizon, driven by recurring glacial bottlenecks, consanguineous inbreeding, and a three-fold relaxation of mitochondrial purifying selection. Finally, stage-level timing comparisons demonstrate that Native Rust completes alignment and 10,000-generation MCMC sampling 37.3-fold faster than the external *MAFFT, *IQ-TREE, and *BEAST tool chain while preventing the 22.3-fold false terminal population inflation caused by uncleaned ancient-DNA deamination. Canonical Preprint URL: https://crvscience.com/labs/preprints/neanderthal-mcmc-skyline-preprint Master (Concept) Zenodo DOI: 10.5281/zenodo.24055714 Open-Source Software Repository (GNU GPL-3.0): https://github.com/bpwhite/crvscience-labs

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-08
DOI
https://doi.org/10.5281/zenodo.24055714
Primary Topic
Forensic and Genetic Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

RuNIE Long-Form Bayesian MCMC Skyline & the 5-Stage Neanderthal Population Crash (650 to 39.2 ka BP): Coupled 95% Credible Chronogram, Pristine Phylogram, and Effective Population Size Trajectory Benchmarked Against *BEAST across an 87,227-bp Alignme

Bryan Pinckney White
3 links
Zenodo (CERN European Organization for Nuclear Research)
Forensic and Genetic Research
preprint

RuNIE Long-Form Bayesian MCMC Skyline & the 5-Stage Neanderthal Population Crash (650 to 39.2 ka BP): Coupled 95% Credible Chronogram, Pristine Phylogram, and Effective Population Size Trajectory Benchmarked Against *BEAST across an 87,227-bp Alignme

Bryan Pinckney White
preprint en

Abstract

Coupled Metropolis-Hastings MCMC sampling (runie-mcmc-skyline: G = 10,000 generations, 20% burn-in = 2,000 discarded, 8,000 posterior states) across 64 ancient/modern hominin mitogenomes and 16 nuclear loci (L = 87,227 *MAFFT / 86,612 Native Rust gotoh-msa aligned bp) resolving the 5-stage Neanderthal Ne(t) collapse (13,893 to 602.7 at 39.2 ka BP) and 37.3x speedup over *MAFFT + *IQ-TREE + *BEAST v1.10.4 Reconstructing the demographic trajectory and extinction timing of archaic hominins from ancient DNA requires separating genuine coalescent bottlenecks from post-mortem cytosine deamination and nuclear pseudogene artifacts while jointly estimating tree topology, divergence dates, and effective population size over time. Here we introduce and validate a configurable long-form Bayesian Markov chain Monte Carlo sampler inside the Rust Native Inference Engine (runie-mcmc-skyline; default 10,000 generations with a 20% burn-in gate) and benchmark it against *BEAST, *IQ-TREE with *LSD2, and *PAML codeml across an 87,227-bp hominin alignment spanning Neanderthals, Denisovans, Sima de los Huesos, and modern humans. Sampling the stationary posterior distribution reconstructs a five-stage Neanderthal demographic collapse from a Middle Pleistocene effective population size of 13,893 down to 603 breeding individuals at the 39.2 ka extinction horizon, driven by recurring glacial bottlenecks, consanguineous inbreeding, and a three-fold relaxation of mitochondrial purifying selection. Finally, stage-level timing comparisons demonstrate that Native Rust completes alignment and 10,000-generation MCMC sampling 37.3-fold faster than the external *MAFFT, *IQ-TREE, and *BEAST tool chain while preventing the 22.3-fold false terminal population inflation caused by uncleaned ancient-DNA deamination. Canonical Preprint URL: https://crvscience.com/labs/preprints/neanderthal-mcmc-skyline-preprint Master (Concept) Zenodo DOI: 10.5281/zenodo.24055714 Open-Source Software Repository (GNU GPL-3.0): https://github.com/bpwhite/crvscience-labs

Zenodo (CERN European Organization for Nuclear Research)
CRV (Netherlands) (NL)
Forensic and Genetic Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon