Searching for the roots of the first free African American community

San Basilio de Palenque, founded by runaway slaves near Cartagena, Colombia, became the first free African community in the Americas by the 17th century. Cartagena was a central Spanish slave-trade port receiving Africans from multiple regions, initially Upper Guinea and later Congo/Angola and broader West–East Africa. Palenque has maintained distinct cultural identity, high endogamy, and the only Spanish-lexifier Creole in the Americas, leading to its UNESCO recognition. Linguistic and anthropological evidence has suggested a primarily Bakongo (Kongo/Angola) origin for founders, potentially a single Bantu group (Kikongo speakers). However, cultural and genetic signals may differ, and prior genetic studies were limited in resolution. Early HLA studies suggested limited Native American and European gene flow and affinity to West African groups; Y-chromosome data indicated substantial European paternal input (~38%) and autosomal AIMs showed ~10% European ancestry, suggesting sex-biased admixture. More recent uniparental work pointed to Yombe (Republic of the Congo) as likely paternal source but left many lineages unresolved, especially mtDNA HVSI-only data. This study aims to refine the continental and regional origins of Palenque’s maternal and paternal lineages, assess Native American and European influx, and test the hypothesis of a single versus multiple African origins using higher-resolution mtDNA control region sequences and Y-chromosomal SNP/STRs.

• Historical records indicate diverse African origins for enslaved people arriving in Cartagena, shifting from Upper Guinea to Congo/Angola and broader regions by the 18th–19th centuries.
• Linguistic/anthropological work posits a Bakongo, Kikongo-speaking origin for Palenque’s founders.
• Genetic studies:
  • HLA markers: limited Native American/European gene flow; closer distances to West African groups.
  • Y-chromosome (earlier work): suggested higher European paternal input (~38%).
  • Autosomal AIMs: ~10% European ancestry, indicating sex-biased admixture.
  • Ansari-Pour et al. analyzed uniparental markers: suggested Yombe as likely paternal source; mtDNA HVSI data left >25% maternal haplotypes with unresolved continental origin; limited resolution outside E1b1a-M2 on Y.
• Gaps: low-resolution markers, incomplete Y-STR/SNP panels, and inability to quantify non-African influx or pinpoint regional African origins across lineages.

• Participants: 95 unrelated male children (5–18 years), self-identified with Palenque ancestry for ≥3 generations (all parents and grandparents born in Palenque). Ethics approvals obtained; guardians provided informed consent. Two subsamples: PR (Palenque residents; n=48) and PU (Cartagena urban schools with Palenque ethno-education program; n=47).
• DNA extraction: standard salting-out from blood.
• Y chromosome:
  • 51 Y-SNPs genotyped (95 samples) using published methods.
  • 27 Y-STR loci (Yfiler Plus) genotyped (92 samples); fragment analysis on 3500 XL; GeneMapper IDX v4.0 for allele calling.
• mtDNA: full control region (positions 16024–576) sequenced in 81 samples using established primers/protocols; haplotypes called with SeqScape v2.7; haplogroups assigned via EMPOP v4/R1 and Haplogrep, confirmed with Phylotree; data QC via EMPOP (EMP00749) and GenBank PopSet 1782793150 (MK930265–MK930345).
• Statistics and comparative analyses:
  • Haplogroup frequencies by direct counting; haplotype/haplogroup diversities, pairwise FST and nondifferentiation probabilities computed in Arlequin v3.5.1.2.
  • MDS visualisation with STATISTICA v8.0.
  • Phylogenetic networks constructed with Network v10.1.0.0; common Y-STR sets (typically 11 loci: DYS389I, DYS389II, DYS19, DYS390, DYS438, DYS392, DYS437, DYS385a/b, DYS393, DYS439; DYS391 added for some clades) to maximise African population coverage.
  • Phylogeographic inference based on haplogroup frequencies, distributions across Africa, and haplotype-level comparisons with literature datasets (Supplementary Tables S4–S5).

• Subsample comparison (PR vs PU):
  • mtDNA haplotypes: no significant differentiation (FST = −0.018; P = 0.9855±0.001).
  • Y-STR haplotypes: Fs = 0.0041; P = 0.1119±0.0014.
  • Y-SNP haplogroups: Fgp = 0.0163; P = 0.0792±0.0012. Pooled for subsequent analyses.
• mtDNA diversity (n=81):
  • Haplogroup diversity: 0.8895±0.0193 (22 haplogroups); haplotype diversity: 0.9225±0.0162 (33 haplotypes; many shared within haplogroups).
  • Haplogroup composition: 91.36% African macro-haplogroup L; 8.64% Native American (A2, A2af1a1, A2a1, B2d, C1c3); no European mtDNA lineages observed.
• Y-chromosome diversity:
  • Y-SNP haplogroup diversity: 0.8185±0.0033 (17 haplogroups); Y-STR haplotype diversity: 0.9881±0.0038 (many shared within haplogroups).
  • Continental attributions: 61.05% African, 35.79% European, 3.16% Native American (Q1a2-M3 x sublineages).
• Genetic structure vs Africa (FST/MDS):
  • Both mtDNA and Y-STR comparisons show large pairwise FST to all African references; MDS plots place Palenque apart from all groups—consistent with strong genetic drift; frequency-based clustering inadequate for pinpointing origins.
• Phylogeographic inferences (selected lineages):
  • mtDNA:
    • L0a1a+200: likely origin between Upper Guinea and Bight of Benin.
    • L1b1a1’4: West Africa (Senegambia/Upper Guinea); L1b1a7a/L1b1a18 widespread/unresolved.
    • L1c3/L1c3a/L1c3a1b: Western/central regions favored; three of five L1c3a1b match Ghana; close to Bantu from Angola and Cameroon.
    • L2a subclades predominant (92% of L2) but too widespread for precise origin; L2a1c3b2 in Palenque differs by 5 sites from sole African comparator.
    • L2b1a: likely western region between Bight of Benin and Angola (excludes East/Southeast); L2d+16129: clear match in Angola (Bantu), indicating Angolan origin.
    • L3d1a1a: 11 samples share single haplotype; origin unresolved, one step from central cluster spanning multiple regions.
    • L3e1d: almost exclusive southwest/south-central distribution; Palenque haplotypes consistent with Angola origin; single founder pattern.
    • L3f1b+16365: closest to Ghana.
  • Y-chromosome:
    • A (Y-MRCA xM13, SRY10831.1): Palenque haplotypes intermediate; exclude East/Southeast/South origins; likely West Africa (Upper Guinea; closest to Guinea-Bissau A1-M31).
    • B2a-M150* (xM109): three samples; profiles compatible with a single founder; closest to Gabon/Cabinda (Loango–Cabinda region; ancient Kingdom of Kongo).
    • E1a-M33: two independent origins in West Africa—Upper Guinea (Guinea-Bissau/Burkina Faso) and Bight of Benin/Nigeria-Benin region.
    • E1b1b-M35* (xM78, M81, M123, V6, M293): likely West Africa between Senegambia and Bight of Benin (though some rare European subclades exist; African origin favored).
    • E1b1a-M2* (xM154, M191): clusters indicate Bight of Benin and regions south of Cameroon (Loango–Angola) origins; multiple founders.
    • E1b1a-M191: matches in Gabon, Angola, Zambia; likely Loango–Angola corridor.
    • R1b-V88: four identical Y-STRs; closest to Punu (Gabon), with ties to Equatorial Guinea and Benin—consistent with western coastal origins.
• Admixture pattern:
  • Strong sex-biased admixture: predominant African maternal ancestry; European admixture mediated by males (≥9 different European Y haplogroups including E1b1b-M81, E1b1b-M123, G, I2-M26, J2-M172, R1a, European R1b clades, etc.).
  • Native American contribution higher maternally than paternally; three Native American Y chromosomes likely a single paternal entry (differences only at rapidly mutating DYF387S1).

The study addressed whether Palenque’s founders derived from a single or multiple African origins and quantified non-African inputs. Frequency-based FST/MDS analyses were confounded by pronounced genetic drift due to long-term isolation, obscuring affinities with specific African regions. A lineage-by-lineage phylogeographic approach instead resolved several maternal and paternal founder origins across multiple West and West-Central African regions—from Upper Guinea through the Bights of Benin and Biafra to Loango and Angola—contradicting a strict single Bakongo (Kongo/Angola) origin. Nevertheless, some lineages do trace to southern West-Central Africa (e.g., Angola), consistent with known historical inputs and cultural dominance of Kikongo. The results corroborate pronounced sex-biased admixture: virtually exclusive African mtDNA with minor Native American maternal input, contrasted with substantial European paternal contribution and minimal Native American Y lineages. These findings align with historical slave-ship routes to Cartagena, recorded from Senegambia to Angola, and illustrate how Palenque’s genetic landscape reflects multiple African substrates shaped by founder effects and drift.

High-resolution uniparental analyses reveal that Palenque’s African lineages descend from a restricted number of founders originating from multiple West and West-Central African regions, not from a single source population. While cultural and linguistic evidence highlights Kikongo influence, genetic data indicate additional African roots. Admixture was strongly sex-biased, with predominant African maternal ancestry and European contributions primarily via males; no European maternal lineages were detected. Future work should (i) refine Y and mtDNA subclade resolution with broader, high-density datasets, (ii) expand sampling in underrepresented African regions relevant to the trans-Atlantic slave trade, and (iii) investigate the historical timing and context of the European paternal legacy in Palenque.

• Strong genetic drift and low haplotype diversity within haplogroups reduce power of frequency-based (FST/PCA) affinity analyses.
• Comparative datasets constrained to HVSI/HVSII for mtDNA (despite full CR sequenced), limiting resolution of phylogeographic inferences.
• Limited overlap in published Y-STR panels and heterogeneous Y-SNP typing schemes across studies hinder fine-scale comparisons.
• Many African regions remain undersampled or analyzed at low resolution, preventing origin assignment for a substantial fraction of lineages and precluding precise estimation of regional contributions.
• Sampling restricted to male children; mtDNA analyzed in a subset (n=81), potentially missing rare lineages.