What the study is

Li and colleagues at Tianjin Medical University published this work in PLoS ONE on September 9, 2026 (DOI 10.1371/journal.pone.0357667). They enrolled 18 patients from the pediatric dentistry department with a radiographically confirmed diagnosis of oligodontia, extracted genomic DNA from peripheral blood, and ran whole-exome sequencing (NovoGene Bioinformatics, Illumina HiSeq 4000). The study was approved by the university’s ethics committee, with recruitment between October 2019 and January 2024. The aim was to find genes associated with tooth agenesis beyond the established list, using the patients who had no mutation in a known causative gene.

What the sequencing found

In 12 of the 18 patients, sequencing found mutations in genes already established to cause tooth agenesis; the paper names WNT10A in three patients (P7, P9, P10), AXIN2 (P8), and FGFR1 (P11), among others. The remaining six patients, numbered 1 through 6, had no mutation in a known causative gene; five of the six were missing anterior teeth and one was missing premolars. These six shared variants in ten genes associated with maxillofacial development: CHD3, EPB41L4A, PRDM2, COL11A2, BPTF, EBF3, IQSEC1, ITPR3, TTN, and WDR33. The detail that drives the paper: all six carried the same EPB41L4A splice-acceptor variant, rs145708081 (NM_022140.5:c.1933-6_1933-2dup).

The cluster argument

Rather than test the ten genes functionally, the authors built a sequence-similarity argument. From NCBI they pulled 20 human mRNA sequences linked to tooth agenesis plus the mRNA sequences of the ten maxillofacial-development genes, ran pairwise BLAST local alignments across the 30 sequences, converted the comparison scores to distances, and clustered twice: hierarchical clustering with Euclidean distance and Ward linkage, and fuzzy K-means clustering. Both methods placed EPB41L4A in the same class as KREMEN1, POLR3A, RNF216, and SEC22B, which the paper describes as genes explicitly associated with congenital tooth loss in the database. Clustered sequences are assumed to share function, and on that assumption the authors nominate EPB41L4A as a plausible candidate gene for tooth agenesis.

Why EPB41L4A is plausible, and what would actually test it

The gene (also called Nbl4) encodes a FERM-superfamily protein that links membrane proteins and lipids to the actin cytoskeleton. Published work cited by the authors supplies the developmental rationale: Ishiguro et al. identified EPB41L4A as a target of the beta-catenin/TCF pathway in mouse and human cells, and Guo et al. reported epb41l4a expression in the cranial skeletal scaffold and other structures during Xenopus embryogenesis. Wnt signaling is one of the two pathways, alongside NF-kB, that dominate the known genetics of tooth agenesis. What is missing is any direct test: no cell assay, no animal knockout, no segregation in families, no replication cohort. The authors state this themselves, twice, noting that the clustering findings require further validation and that their analysis stops at the nucleotide-sequence level.

Where the honest boundary sits

Clustering BLAST similarity scores is a long inferential chain: sequence similarity is taken to imply functional similarity, which is taken to imply disease involvement. That can prioritize a shortlist; it cannot show that rs145708081 causes a missing tooth. The paper does not establish how common this variant is in the general population, so even the observation that all six unexplained patients carry it lacks a frequency denominator, and population frequency is exactly what a splice-region variant needs before it can be read as causal. The cohort is one center and 18 patients, of whom only six drive the EPB41L4A claim. The correct reading is a candidate gene nomination, nothing more.

Why it matters to the third dentition field

Congenital agenesis is the named indication of the only whole-tooth program in human trials, the anti-USAG-1 antibody TRG-035 (program page), which holds Japanese orphan drug designation for severe congenital partial anodontia. Agenesis is genetically heterogeneous, and which gene is broken in a given patient plausibly shapes whether relieving USAG-1-mediated suppression can rescue tooth formation at all. If EPB41L4A holds up, it extends the map of pathways that must fail for a tooth not to form, and its Wnt connection sits in the same signaling territory that USAG-1 inhibition acts on, since USAG-1 (SOSTDC1) antagonizes both BMP and Wnt. That is a connection of shared pathway, not of evidence: this paper tests no treatment, advances no regeneration result, and touches no program tier. The current field assessment stands at /field/.

Provenance: grounded in the full open-access text of Li et al. 2026 (PLoS ONE 21(9):e0357667), read in full, with program context from /programs/anti-usag-1-trg035/. Method and sourcing standard at /method/.