Ancestry and Population Affinity from Dentition

The morphological features of tooth crowns and roots are among the most heritable of all skeletal traits, with heritability estimates for individual traits ranging from roughly 0.5 to 0.9. Unlike long-bone dimensions, which are heavily influenced by nutrition, disease, and physical loading during life, tooth shape is largely fixed during crown formation in childhood and does not remodel. This makes dental morphology a particularly stable genetic signal.

The evolutionary context matters. When human populations colonised different parts of the world, founding bottlenecks and subsequent genetic drift caused certain trait frequencies to diverge. Shovelling, for instance, reached very high frequency in the East Asian and Beringian populations that eventually crossed into the Americas, possibly because a variant of the EDARV370A gene that drives shovelling also conferred other advantages in those environments. That frequency contrast between Northeast Asian and sub-Saharan African populations is real, large, and detectable in a forensic case.

The caution is that no trait is absent in any living population, only rarer. A forensic analyst working from trait frequencies never asks 'does this person have trait X?' and concludes a group membership. The question is always: given this pattern of trait frequencies across multiple traits, which reference population is the best probabilistic match? That is a quantitative question, not a typological one.

Christy Turner II at Arizona State University spent decades from the late 1960s through the 1990s systematically recording dental morphology across global skeletal collections. The problem he addressed was that researchers were using their own non-standardised trait descriptions, making comparisons across studies nearly impossible. ASUDAS introduced physical reference plaques for each trait, fixing the grade boundaries so that an analyst in Tokyo and an analyst in London looking at the same tooth could arrive at the same score.

The current system covers 29 traits across the crown and root. Selected traits of greatest forensic relevance include:

Shovelling refers to raised lingual marginal ridges on the upper central incisors, and to a lesser degree the upper lateral incisors and lower incisors, visible when the tooth is viewed from the tongue side. ASUDAS grades run from 1 (faint trace) to 7 (pronounced bilateral ridges). A semi-shovel threshold at grade 3 and full shovel at grade 4 and above are the most commonly reported levels in the forensic literature.

Genome-wide association studies and functional work have linked high-grade shovelling to a derived variant of the EDAR gene (EDARV370A, rs3827760) that reached near-fixation in East Asian and Indigenous American populations. The same variant affects breast morphology, sweat gland density, and hair shaft thickness. Its broad phenotypic effects suggest selection for one or more of these associated traits rather than for shovelling itself, but the dental frequency contrast it produces is among the most pronounced population-level signals in the human skeleton.

Forensic significance: when grade 4 or higher shovelling is present in both upper central incisors, it substantially increases the probability that the individual has East Asian or Indigenous American ancestry. But the calculus is probabilistic. A European individual with grade 3 shovelling exists. A trait-positive result shifts the probability distribution; it does not define group membership.

The Carabelli trait is an accessory cusp or groove on the mesiolingual surface of the upper first molar, occasionally extending to the second molar. Its ASUDAS grades run from 1 (small groove) through 7 (large free cusp). It is more common in European, Middle Eastern, and some South Asian populations, with frequencies for distinct expression (grades 4 to 7) ranging from about 20 to 60 percent in many European samples, compared to roughly 5 to 15 percent in East Asian samples.

Other diagnostically useful traits include the three-rooted lower first molar. Most lower first molars have two roots (one mesial, one distal). A proportion of individuals, more often of East Asian ancestry, have an additional distolingual root. In some Northeast Asian populations this reaches 35 to 40 percent; in European populations it is typically below 5 percent. When present, this trait strongly suggests East Asian ancestry, though its absence cannot rule it out.

• Cusp 5 on the lower first molar (distal accessory cusp): more frequent in sub-Saharan African and some Pacific populations; useful when combined with other trait evidence.
• Cusp 6 (entoconulid): relatively higher in African and some Pacific populations; evaluated on the lower molars with ASUDAS grades 1 to 5.
• Protostylid (labial accessory cusp, lower molars): higher in some Indigenous American and East Asian groups; one of the more variable cross-population traits.
• Peg or reduced upper lateral incisor: more common in East Asian populations, less common in African ones; often evaluated alongside shovelling.

Once ASUDAS scores are recorded for an unknown individual, the analyst needs a way to translate the trait profile into a population affinity estimate. Several approaches exist, ranging from visual comparison against reference frequency tables to formal probabilistic models.

1. Frequency table comparison
   The simplest approach: compare each trait score against published frequency tables for major population groups (East Asian, European, sub-Saharan African, Indigenous American, etc.) and note which group's expected frequencies best match the observed scores. Informal but useful as a first pass.
2. Discrete traits likelihood ratio
   For each trait, compute the likelihood ratio: frequency of the observed score in reference group A divided by frequency in reference group B. Multiply across independent traits (with caution about trait interdependence). A product above a threshold supports group A; below supports group B.
3. MCLUST or Bayesian discriminant methods
   Software tools built specifically for discrete dental trait data apply Bayesian posterior probability estimation or discriminant function analysis adapted for ordinal data. Output is a probability distribution across reference populations rather than a single classification.
4. Integrated skeletal and dental analysis
   In a full forensic biological profile, dental population affinity is combined with craniometric data, postcranial morphology, and any available geographic or contextual information. The combined estimate is more reliable than any single line of evidence.

Ancestry estimation in forensic anthropology involves genuine scientific and ethical tensions, and dental methods are no exception. Three distinct concerns warrant separation, because conflating them produces unreliable science and legally indefensible reports.

The first concern is accuracy. ASUDAS-based population affinity estimates, when carefully done with well-matched reference populations, can narrow the search in an unidentified persons case. They are not definitive, but they are informative. The accuracy claim should cite the reference population and the method used. No forensic report should read 'this individual is East Asian.' The appropriate form is 'the dental morphological profile is most consistent with an East Asian ancestry, with a posterior probability of X based on reference sample Y.' That framing is both honest and legally defensible.

The second concern is the difference between biological population affinity and social race. Legal and census categories for race vary by country and change over time. The dental traits in ASUDAS reflect genetic drift and population history over thousands of years, not the social categories in a missing persons database. An analyst who translates a dental affinity estimate directly into a racial category without acknowledging the gap between these two frameworks is doing the case a disservice.

The third concern is misuse. Ancestry estimates have historically been used in ways that caused harm, including to justify exclusion or to assign stereotyped characteristics beyond what the biological data supports. A forensic odontologist's responsibility ends at the biological profile: most likely ancestry from a constrained list of reference populations, with stated confidence. Inferences about culture, language, behaviour, or social identity are not part of the scientific output.

The goal of ancestry estimation in forensic anthropology is investigative: to narrow a search, not to categorise a person.