The literature

This page exists so that nobody has to take the article on trust.  Every source it rests on is listed here with what it actually measured, who wrote it, and which way it cuts.

It is deliberately not a reading list of things that agree with this site.  The source that helps the young-earth position most is Amos and Hoffman 2010, a mainstream paper that finds two discrete bottlenecks where the standard model assumes a hundred.  The most honest source on the creationist side is Carter 2024, who rejects the popular Babel geography in print and says the Y tree does not display three founding branches.  Both are tagged by what their data show, not by what their authors believe.

How to read an entry

Measurement or model.  A count of alleles in a genotype file is an observation.  An origin location fitted by regression is a model, and it inherits every assumption in the model that produced it.  Both are useful and they are not the same kind of thing.  The origin scans in this article are models.  The heterozygosity, allelic richness and F_ST tables are observations.

Direction.  Assigned by what the source shows, not by what its author believes.

Verified.  Whether this project read the text directly (full), only its abstract or supplement (abstract), or only through another source that quotes it (via).  Nothing is quoted in the main article unless it is marked full or abstract, and every quotation was read out of downloaded bytes.  No quotation, page number or numeral in this article was established with an automated summarising fetcher.

Field definitions

Coverage and its limits, stated plainly.  This catalogue covers the origin-scan literature from Prugnolle 2005 to Henn 2011, the archaic-admixture papers that bear on dispersal geography, and every creationist source found anywhere that addresses African genetic diversity.  Two items the first draft could not obtain were obtained by the verifier and are listed at the bottom;  both are now cited.

The gradient itself

Geography predicts neutral genetic diversity of human populations

Prugnolle, Franck, Andrea Manica & François Balloux.  Current Biology 15(5):R159–R160, 2005.  https://doi.org/10.1016/j.cub.2005.02.038

camp: mainstream topic: gradient method: measurement direction: undercuts verified: full

Key figures: 51 populations, 377 Marshfield-panel microsatellites;  R-squared 85% for heterozygosity against land-routed distance from an assumed Ethiopian origin;  P < 10⁻⁴;  no detectable effect of historical census size.

The earlier and simpler version of the Ramachandran analysis, and the source of the single most useful sentence in this article for testing the young-earth position.  Prugnolle and colleagues looked for a census-size effect on diversity in these data and could not find one, which is the published test of the "Africa grew biggest so it kept the most" rescue.  No creationist source located for this article cites it.  The paper also reads its own result as "a very large number of small bottlenecks", which Amos and Hoffman contradict on comparable data five years later.  Obtained through a headless-browser fetch of the publisher PDF after curl was refused with a 403;  three-column layout, quotations re-extracted column by column.  Cited.

Support from the relationship of genetic and geographic distance in human populations for a serial founder effect originating in Africa

Ramachandran, Sohini, et al.  PNAS 102(44):15942–15947, 2005.  https://doi.org/10.1073/pnas.0507611102

camp: mainstream topic: origin-scan method: model direction: undercuts verified: full

Key figures: 1,027 individuals, 53 populations, 783 autosomal microsatellites;  F_ST against distance R-squared 0.5882 plain and 0.7834 waypoint-routed;  origin scan over 4,210 lattice points;  African origins R-squared 0.757 to 0.870, non-African 1.67 × 10⁻⁷ to 0.744;  best origin 4.3N, 12.8E.

The central source of the article, and the reason the question is decidable at all.  Its scan over every candidate origin on Earth turns a description into a test, and its genotypes and coordinates are posted publicly, which is what made the refit in Part 2 possible.  Two things in it are quoted against the naive use of its own result:  the authors state that this is a serial founder model rather than isolation by distance at equilibrium, and they warn that thin African sampling limits how precisely the origin can be localised.  One published number, the Figure 4A regression from Addis Ababa, does not reproduce under this project's routing;  the scan statistics reproduce to three decimals.  Cited heavily.

Worldwide Human Relationships Inferred from Genome-Wide Patterns of Variation

Li, Jun Z., et al.  Science 319(5866):1100–1104, 2008.  https://doi.org/10.1126/science.1153717

camp: mainstream topic: gradient method: measurement direction: undercuts verified: full

Key figures: 938 individuals, 51 HGDP-CEPH populations, about 650,000 SNPs;  haplotype heterozygosity against distance from Addis Ababa r = −0.91, slope −1.1 × 10⁻⁵ per km;  average allele frequency flattening from 0.04 in sub-Saharan Africa to 0.02 in Oceania and the Americas.

Closes the marker-type escape route.  The same panel of populations, a completely different marker system, and essentially the same correlation:  r = −0.91 is R-squared 0.83 against the microsatellite figure of about 0.86.  A reader who suspects that microsatellite mutation dynamics generate the cline has to explain the SNP haplotypes too, and the ascertainment bias in a SNP array chosen on European and Asian cohorts works against the African signal rather than for it.  science.org refused both curl and a headless browser;  obtained as an open PDF from a university course archive.  Cited.

Evidence that two main bottleneck events shaped modern human genetic diversity

Amos, William & Joseph I. Hoffman.  Proceedings of the Royal Society B 277(1678):131–137, 2010.  https://doi.org/10.1098/rspb.2009.1473

camp: mainstream topic: gradient method: measurement direction: complicates verified: full

Key figures: the same 783 microsatellites in the same 53 populations Ramachandran used, analysed with BOTTLENECK;  two primary events, one leaving Africa and one at the Bering Strait;  strongest bottleneck signal in the Middle East, weakest in Africa.

The most useful mainstream source in this dossier for the young-earth side.  The serial founder model Ramachandran fitted assumes 100 successive bottlenecks of 250 founders.  Amos and Hoffman, on the identical dataset, find the loss concentrated in two events and say in print that this contrasts with the smooth gradient the standard plot suggests.  A dispersal with a small number of large founder events is closer in shape to a Babel story than the cline picture is.  Then read the location of the events:  the Middle East carries the founder signature and Africa carries none, which is the reverse of what the Babel model needs.  Open author PDF;  folios checked on every page.  Cited.

Where in Africa, and how firmly

The Genetic Structure and History of Africans and African Americans

Tishkoff, Sarah A., et al.  Science 324(5930):1035–1044, 2009.  https://doi.org/10.1126/science.1172257

camp: mainstream topic: origin-scan method: measurement direction: undercuts verified: full

Key figures: 1,327 markers, 121 African populations, 2,432 Africans plus 60 non-African populations;  variation among African populations 1.71% against 8.36% among Native American and 4.59% among Oceanic;  origin fitted at 12.5°E, 17.5°S with a Red Sea waypoint;  14 ancestral clusters at K = 14.

Answers the thinnest part of the HGDP-CEPH panel, which is African sampling density, and finds the same pattern:  highest within-population diversity, declining diversity with distance, more private alleles in Africa.  The 1.71% against 8.36% comparison is an independent version of the F_ST table computed in Part 4 of the article, in different units and on a different sample, and it points the same way.  Its own origin estimate lands 4,000 km from Ramachandran's, which is reported in the article as a genuine weakness of all such scans.  Author manuscript from PMC.  Cited.

Hunter-gatherer genomic diversity suggests a southern African origin for modern humans

Henn, Brenna M., et al.  PNAS 108(13):5154–5162, 2011.  https://doi.org/10.1073/pnas.1017511108

camp: mainstream topic: origin-scan method: model direction: undercuts verified: full

Key figures: more than 580,000 SNPs in Hadza, Sandawe and ≠Khomani Bushmen plus 24 comparison populations;  LD-based origin 14S, 12E with r = 0.78;  F_ST-based origin 20S, 22E with r = −0.84 and P = 0.06;  ≠Khomani HLA-A heterozygosity 0.95 across 30 alleles.

Repeats the origin scan on two statistics that admixture cannot inflate, which is the obvious objection to a heterozygosity-based scan, and gets the same continent by both.  It is also the honest place to record how weak the within-Africa resolution is:  its own F_ST result reaches only P = 0.06, and its estimate is 4,000 km from Ramachandran's.  Its HLA numbers, 30 alleles in one small hunter-gatherer group, are the diversity figures hardest for any recent-founder model to produce.  Cited for the LD and F_ST scans and for the P value;  the HLA measurement is left to the population bottlenecks article, which handles allele counts at classical loci.

Human origins in Southern African palaeo-wetlands? Strong claims from weak evidence

Schlebusch, Carina M., et al., including Chris Stringer, Mattias Jakobsson, Andrea Manica, Sarah Tishkoff, Eleanor Scerri, Aylwyn Scally and Mark Thomas.  Journal of Archaeological Science 130:105374, 2021.  https://doi.org/10.1016/j.jas.2021.105374

camp: mainstream topic: critique method: commentary direction: complicates verified: full

Key figures: a multi-author rejection of a Nature paper claiming that mitochondrial DNA localises human origins to a specific southern African wetland.

A large group of the field's own leading names publicly rejecting a high-profile origin claim on methodological grounds:  one locus carries one genealogy and cannot localise a population's origin, and present-day distributions are poor guides to past ones.  The same objection applies in weaker form to every origin scan in this article, including the reproduction in Part 2, and the article says so.  This is why no birthplace is claimed here.  Cited.

What the creationist literature says

The Neutral Model of evolution and recent African origins

Carter, Robert W.  Journal of Creation 23(1):70–77, 2009.

camp: creationist topic: african-diversity method: hypothesis direction: complicates verified: full

Key figures: the only published creationist mechanism for the African diversity excess:  small, isolated, structured post-Flood African tribes with inbreeding, rapid drift and rapid mutation accumulation.

The most substantial creationist engagement with African genetic diversity that this project's search located, and a real critique of the Standard Neutral Model's assumptions.  Its positive mechanism predicts the opposite of the observation.  Small isolated populations lose within-population heterozygosity, and the American row of the measured table shows exactly that pattern:  the highest F_ST and the lowest heterozygosity in the world.  Africa shows the reverse.  Carter's third route, admixture, is the only fast way to raise heterozygosity, and it would lower F_ST among the mixing groups rather than leave African F_ST second-highest.  He does not cite Ramachandran, Prugnolle, Handley or Manica.  He does cite Liu 2006, as his reference 24, in a single clause about model complexity that says nothing about what Liu found.  Two-column PDF, quotations re-extracted with cropping, folios anchored twice.  Cited.

Genetic diversity in humans

Carter, Robert W.  Creation Ministries International, 2011.

camp: creationist topic: yec-argument method: hypothesis direction: complicates verified: full

Key figures: the four-allele model:  Adam and Eve could supply up to four versions at a locus, and third, fourth and fifth versions are post-Babel mutations restricted to subpopulations.

The clearest statement of the creationist expectation for allele counts, and the one most directly checkable against the Ramachandran genotypes.  Checked in Part 5:  mean 11.93 distinct alleles per locus, maximum 35, and 428 of 783 loci above the post-Flood ceiling of ten.  That result is not fatal on its own because microsatellites mutate fast, and the article says so.  The claim fails on the geography of the extra alleles rather than their number.  Carter's opening concession, that worldwide human diversity is low and the mainstream reads it as a bottleneck, is quoted in the article for balance because it is correct.  Cited.

African Populations Fit Biblical History

Thomas, Brian.  Institute for Creation Research, 2012.

camp: creationist topic: african-diversity method: commentary direction: undercuts verified: full

Key figures: a post-Babel reading of Lachance et al. 2012:  intermarriage at Babel, dispersal, Ice Age isolation in Africa, recent diversification;  quotes 13,407,517 SNPs from that study.

One of only three creationist sources anywhere that addresses African genetic diversity.  Two of its steps run against the argument it defends.  Ice Age isolation lowers within-population heterozygosity, which is the quantity that has to come out high.  And the SNP count it quotes approvingly is a cost under its own model, since every variant beyond ten per site has to be generated in about 160 generations.  Its source, Lachance et al. 2012, has now been read:  Lachance counts 13,407,517 variants, meaning SNPs, insertions and deletions together across fifteen genomes, where ICR calls all of them SNPs.  Cited, with the correction.

On the Origin of Human Mitochondrial DNA Differences, New Generation Time Data …

Jeanson, Nathaniel T.  Answers Research Journal 9:123–130, 2016.

camp: creationist topic: african-diversity method: hypothesis direction: undercuts verified: full

Key figures: a proposed historical African generation interval of 15 years, derived from UN data on the percentage of women married by age 19.

The more specific of the two creationist mechanisms, and it deserves a fair hearing because the arithmetic works.  Extended from mitochondria to microsatellites, 107 extra generations in 4,000 years buys about 0.064 of heterozygosity under the infinite-alleles bound, against an observed African excess of 0.045.  The mechanism is numerically capable.  It fails on its input:  Wang 2023 measures African generation intervals as the longest of any continental group, not the shortest.  Two further problems are that percentage married by 19 is not a generation interval, and that a mechanism working through generations elapsed predicts no linear relation with kilometres.  Cited.

On the Origin of Human Mitochondrial DNA Differences … (generation-time input)

Wang, Richard J., Samer I. Al-Saffar, Jeffrey Rogers & Matthew W. Hahn.  "Human generation times across the past 250,000 years."  Science Advances 9(1):eabm7047, 2023.  https://doi.org/10.1126/sciadv.abm7047

camp: mainstream topic: african-diversity method: measurement direction: undercuts verified: full

Key figures: generation intervals read out of the mutation spectra of 1,247 sequenced trios plus dated variants;  East Asian 20.1 ± 3.9, European 20.6 ± 3.8, South Asian 21.0 ± 3.7, African 26.9 ± 3.5 years, P < 1 × 10⁻¹⁰.

The single measurement that decides Jeanson's proposal, and it decides it against.  Africans have the longest measured sex-averaged generation interval of any continental group, by about six years, which is the reverse of the direction the mechanism requires.  It is a genetic measurement rather than a demographic one, so it reaches back thousands of generations rather than to the era of civil registration.  In the source library;  the page was anchored from the running foot and the n of 5 page mark.  Cited.

Out of Babel — not Africa: genetic evidence for a biblical model of human origins

Tomkins, Jeffrey P.  Journal of Creation 34(1):79–85, 2020.

camp: creationist topic: yec-argument method: review direction: undercuts verified: full

Key figures: four arguments, none of which touches the heterozygosity gradient.

The flagship creationist statement of the exact position this article examines, and the word heterozygosity does not appear in it.  Neither does any origin scan, nor Ramachandran, Prugnolle, Liu, Handley, Manica or Li.  It does cite Henn, Steele and Weaver 2018, and Carter 2009.  Its archaic-admixture argument also overstates its own source:  it presents Chen et al. 2020 as a wrench in the Out of Africa paradigm, states that paper's own back-migration explanation in one sentence, and then changes the section without answering it.  The Henn, Steele and Weaver point that four sub-models of African origins compete is fair and is reported in the article, with the note that it is a dispute about where inside Africa.  Cited.

Can we place the sons of Noah on the Y chromosome tree?

Carter, Robert W.  Creation Ministries International, 29 October 2024.

camp: creationist topic: tree-shape method: commentary direction: complicates verified: full

Key figures: rejects the Japheth-Europe, Ham-Africa, Shem-Middle-East mapping;  concedes the Y tree does not display three founding branches.

The fairest creationist source in this dossier.  Carter dismantles the popular Babel geography himself, in print, and calls it "entirely wrong".  His concession about the tree's shape is quoted in the article.  The cost he does not price is that once the three sons' descendants overlap everywhere, the Table of Nations stops making any checkable prediction about the geography of Y haplogroups.  Cited for the geography;  the Y-tree topology is handled in Y-chromosome Adam.

Archaic admixture and dispersal shape

The Date of Interbreeding between Neandertals and Modern Humans

Sankararaman, Sriram, Nick Patterson, Heng Li, Svante Pääbo & David Reich.  PLoS Genetics 8(10):e1002947, 2012.  https://doi.org/10.1371/journal.pgen.1002947

camp: mainstream topic: archaic-admixture method: measurement direction: undercuts verified: full

Key figures: 1,805–1,993 generations since Neanderthal gene flow into Europeans from the deCODE pedigree map, 1,881–2,043 from the European LD map, both 95% credible intervals;  37,000–86,000 years after conversion.

The strongest single argument in the article, because it needs no calendar.  Admixture linkage disequilibrium decays with each meiosis, and the deCODE map is built by watching crossovers in living pedigrees, so the primary output is a count of generations.  A 4,000-year post-Babel window is 160 generations at 25 years each.  The measurement is nine to twelve times larger.  Closing the gap needs a two-year generation interval, an unmeasured ancient recombination rate, or a non-admixture source for the Neanderthal-matching haplotypes, and this paper's test was designed to distinguish admixture from ancient structure.  Open access, cc-by.  Cited.

Identifying and Interpreting Apparent Neanderthal Ancestry in African Individuals

Chen, Lu, Aaron B. Wolf, Wenqing Fu, Liming Li & Joshua M. Akey.  Cell 180(4):677–687.e16, 2020.  https://doi.org/10.1016/j.cell.2020.01.012

camp: mainstream topic: archaic-admixture method: measurement direction: complicates verified: full

Key figures: 17 Mb of Neanderthal sequence per African individual by IBDmix against under 0.5 Mb by earlier methods;  51 Mb in Europeans, 55 Mb in East Asians and South Asians;  1.2 Mb of Denisovan sequence in Africans against about 1 Mb in non-Africans.

The source Tomkins 2020 leans on hardest, so the article states what it actually says.  Its own explanation for African Neanderthal ancestry is back-migration of Eurasians into Africa plus early human-to-Neanderthal gene flow, printed in its section heading, and it presents itself as a refinement of Out of Africa.  The genuinely useful lesson is methodological and the article gives it plainly:  a decade of studies reported no Neanderthal ancestry in Africans because the method used Africans as the unadmixed reference.  Cited.

The Simons Genome Diversity Project: 300 genomes from 142 diverse populations

Mallick, Swapan, et al. and David Reich.  Nature 538(7624):201–206, 2016.  https://doi.org/10.1038/nature18964

camp: mainstream topic: archaic-admixture method: model direction: complicates verified: full

Key figures: 300 genomes at 34–83-fold coverage from 142 populations;  no position or proportion of a deeply diverging modern human lineage improves the model fit for Australians, New Guineans or Andamanese.

Cited in the article as one half of a live mainstream dispute, alongside Pagani 2016 in the same issue of Nature.  A single-source dispersal is what a Babel model wants too, so this is not by itself a mainstream-versus-Babel result;  it bites because the single source is placed on the African side of the split.  Cited.

Genomic analyses inform on migration events during the peopling of Eurasia

Pagani, Luca, et al. and Mait Metspalu.  Nature 538(7624):238–242, 2016.  https://doi.org/10.1038/nature19792

camp: mainstream topic: archaic-admixture method: model direction: complicates verified: full

Key figures: 483 high-coverage genomes from 148 populations;  at least 2% of the Papuan genome from an earlier, extinct dispersal;  split dated about 120,000 years ago.

The other half of the dispute.  Two papers on comparable data in one issue of Nature, one finding no early-dispersal ancestry anywhere and the other finding at least 2% of it in Papuans.  The article reports the dispute without inflating it and states the distance:  2% of one population's genome, and 120,000 years against 60,000, with neither position within four orders of magnitude of a post-Babel dispersal.  Cited.

A recent bottleneck of Y chromosome diversity coincides with a global change in culture

Karmin, Monika, et al.  Genome Research 25(4):459–466, 2015.  https://doi.org/10.1101/gr.186684.114

camp: mainstream topic: tree-shape method: measurement direction: complicates verified: full

Key figures: Y MRCA in Africa at 254 kya (95% CI 192–307);  a cluster of major non-African founder haplogroups at 47–52 kya.

Reused from the Y-chromosome Adam dossier for the geography only.  The founder-haplogroup cluster is the shape a Babel dispersal predicts, a burst of founder lineages in a narrow interval, and the article says so.  The disagreement is entirely about what sits beneath it, since the cluster is below four African-only branches.  Cited once.

A 'Copernican' Reassessment of the Human Mitochondrial DNA Tree from its Root

Behar, Doron M., et al.  American Journal of Human Genetics 90(4):675–684, 2012.  https://doi.org/10.1016/j.ajhg.2012.03.002

camp: mainstream topic: tree-shape method: measurement direction: undercuts verified: full

Key figures: 18,843 mitogenomes;  mean root-to-tip distance 57.1 substitutions, range 41 to 77;  39 substitutions from the reconstructed root to the root of M.

Cited once, for the fact that the whole non-African world descends from two twigs of L3, which itself sits 39 substitutions below the root.  Under an out-of-Mesopotamia dispersal the deepest branches should sit near Mesopotamia and Africa should hold a derived subset.  The observed pattern is the reverse.  The two-child structure of the mitochondrial root is established in agents/memory/mitochondrial-eve.md from PhyloTree Build 17 with a Neanderthal outgroup, not asserted from this paper.  The full argument is in mitochondrial Eve.

The data

HGDP-CEPH genotypes used by Ramachandran et al. 2005

Ramachandran, Sohini & Noah A. Rosenberg (Rosenberg lab release).  783 autosomal microsatellites × 1,027 individuals, plus the 53-population coordinate and code files.

camp: mainstream topic: gradient method: measurement direction: undercuts verified: full

Key figures: 6.4 MB genotype file;  mean 11.93 distinct alleles per locus, maximum 35, 428 of 783 loci above ten alleles;  African private allelic richness 0.168 per locus against 0.017 to 0.031 elsewhere at matched sample size.

The raw material for everything computed in the article.  Posted publicly by the Rosenberg lab without a stated licence, so the library records it as unknown.  Two Deno programs in research/code/ read it:  origin-scan.ts recomputes heterozygosity, routes the distances, refits the regression at any candidate origin, runs the lattice scan and the non-African refit, and bootstraps;  richness.ts computes rarefied allelic richness, private allelic richness and within-region F_ST.  Both were re-run for this article and reproduce their stored outputs byte for byte.

Considered and not used

Genetic Structure of Human Populations

Rosenberg, Noah A., et al.  Science 298(5602):2381–2385, 2002.  https://doi.org/10.1126/science.1078311

camp: mainstream topic: gradient method: measurement direction: complicates verified: full

Considered and not used in the article.  The parent study of the dataset, and the paper most often misused on both sides:  93 to 95% of variation is within populations, and the small remainder still lets a clustering algorithm recover continental groups from genotypes alone.  Both halves are true and both are routinely quoted without the other.  It is a structure paper rather than an origin paper, and the origin scans carry the argument, so citing it would have added length without adding a test.  Its one load-bearing point, that the first split is Africa against everyone else rather than a three-way split, is made more directly by Behar's mitochondrial tree.

Denisova Admixture and the First Modern Human Dispersals into Southeast Asia and Oceania

Reich, David, et al.  American Journal of Human Genetics 89(4):516–528, 2011.  https://doi.org/10.1016/j.ajhg.2011.09.005

camp: mainstream topic: archaic-admixture method: measurement direction: complicates verified: full

Considered and not used in the article.  Denisovan ancestry at 4 to 6% in Near Oceanians and statistically zero in mainland East Asians a few thousand kilometres away.  Sharply patchy archaic ancestry is what a dispersal with limited local contacts looks like, and a Babel model can tell that story as easily as an Out of Africa model can, so the source does not discriminate between them.  What it would add, that the Neanderthal fraction is nearly uniform across Europeans, East Asians and Native Americans while the Denisovan fraction is confined to one branch, belongs to Neanderthals and Denisovans and is argued there.

Multiple Deeply Divergent Denisovan Ancestries in Papuans

Jacobs, Guy S., et al.  Cell 177(4):1010–1021.e32, 2019.  https://doi.org/10.1016/j.cell.2019.02.035

camp: mainstream topic: archaic-admixture method: measurement direction: complicates verified: full

Considered and not used in the article.  161 new genomes across 14 island groups;  Papuans carry haplotypes from two Denisovan lineages that separated more than 350,000 years ago, with one introgression as late as 30 kya.  The 350,000-year separation is a real cost to a young model, but it is a cost about archaic populations rather than about dispersal geography, and it is argued in Neanderthals and Denisovans.  Citing it here would have duplicated a sibling article.  Obtained through a headless-browser fetch of the publisher PDF after curl was refused.

Altitude adaptation in Tibetans caused by introgression of Denisovan-like DNA

Huerta-Sánchez, Emilia, et al.  Nature 512(7513):194–197, 2014.  https://doi.org/10.1038/nature13408

camp: mainstream topic: archaic-admixture method: measurement direction: complicates verified: full

Considered and not used in the article.  A 32.7 kb EPAS1 haplotype carried by nearly all Tibetans and matching the Denisovan genome, with ancestral lineage sorting estimated at p = 0.0012.  On the young-earth side this is easier to accommodate than it looks, since intermarriage between neighbouring people groups produces the same pattern.  It bears on archaic identity rather than on dispersal geography, and it belongs to Neanderthals and Denisovans.

Ancient bottlenecks and phenotypic variation

Manica, Andrea, et al., 2007, and Handley et al. 2007.

camp: mainstream topic: gradient method: measurement direction: undercuts verified: abstract

Considered and not used in the article.  Named in the article only in the list of origin-scan literature that no creationist source cites.  The phenotypic-variation extension of the gradient argument is a separate question from dispersal geography, and the three scans already cited carry the point.

Obtained on the verifier's second pass

Both items the first draft recorded as unobtainable were free at Europe PMC.  The lesson is the standing one:  "I could not find it" is not "it does not exist."

A Geographically Explicit Genetic Model of Worldwide Human-Settlement History

Liu, Hua, Franck Prugnolle, Andrea Manica & François Balloux.  American Journal of Human Genetics 79(2):230–237, 2006.  doi:10.1086/505436.  PMC1559480.

Measurement or model: fits a stepping-stone colonisation model, with coalescence times rather than a regression, to the same 783 microsatellites in 52 HGDP populations.  Gives an initial expansion 56,063 ± 5,678 years ago from a founding effective population of 1,064 ± 93, a deme carrying capacity of 751 ± 155, and about 23% of individuals moving to an adjacent population per generation.

Direction: complicates both sides.  Verified: full.

Access: curl -sL on https://europepmc.org/articles/PMC1559480?pdf=render returns the publisher PDF with no login.  Read with pdftotext -raw, because the two-column layout interleaves under -layout.  In the source library as liu-2006.

It is not a third origin scan, which is what the first draft assumed.  The origin is assumed, not fitted:  Figure 2 is captioned "Shortest colonization routes from East Africa (Addis Ababa) along landmasses".  Its real value here is a warning on p. 234 that cuts against this article's own method, footnoted to Ramachandran 2005:  "An important corollary stemming from this nonlinearity is that use of linear regressions on gene diversities might not allow correct inference of the geographic origin of modern humans."  Their expected diversity is non-linear in distance because populations in the middle of a chain of stepping stones receive migrants from both sides.  Carter 2009 lists this paper as his reference 24 and mentions it in one clause without engaging its result.  Cited in Part 3 and in "Issues and unanswered questions".

Evolutionary History and Adaptation from High-Coverage Whole-Genome Sequences of Diverse African Hunter-Gatherers

Lachance, Joseph, Benjamin Vernot, Clara C. Elbers, et al. and Sarah A. Tishkoff.  Cell 150(3):457–469, 2012.  doi:10.1016/j.cell.2012.07.009.  PMC3426505.

Measurement: high-coverage genomes of 15 African hunter-gatherers (Hadza, Sandawe, Pygmy), plus two Hadza technical replicates.  "After applying stringent quality control filters … we identified 13,407,517 variants (SNPs, insertions, and deletions that differ from the human genome reference sequence, GRCh37/hg19) in the 15 hunter-gatherer genomes."

Direction: undercuts the young-earth argument that quotes it.  Verified: full.

Access: bin/pmcget.py on https://pmc.ncbi.nlm.nih.gov/articles/PMC3426505/, tags stripped locally.  In the source library as lachance-2012.  The ?pdf=render route that worked for Liu 2006 returns a 500 for this item, and cell.com 403s.

Obtaining it settles the one open flag in the article:  the 2012 ICR piece calls all 13,407,517 of them SNPs, where Lachance counts SNPs, insertions and deletions together across fifteen genomes.  The restriction on attributing anything to Lachance is lifted, and the article now cites the sentence directly.  Cited.