Type 2 diabetes mellitus (T2D) is a heterogeneous disease, with certain cases presenting an autosomal dominant type. The rare coding variants of disease‑causing genes in T2D remain mostly unclear. The present study aimed to identify the disease‑causing gene conducting whole exome sequencing in a Thai T2D family with an autosomal dominant transmission of T2D with no evidence of mutations in known maturity‑onset diabetes of the young (MODY) genes. Candidate variants were selected according to certain criteria of mutation prediction programs, followed by segregation analysis with diabetes in the family. The results demonstrated that, of the 68,817 variants obtained, 122 were considered as candidate variants subsequent to the filtering processes. Genotyping of these variants revealed that DnaJ homolog subfamily C member 3 (DNAJC3) p.H238N segregated with diabetes in the family. This mutation was also identified in another proband from the autosomal dominant T2D family without mutation in known MODY genes and was segregated with diabetes. This variant was also identified in 14/1,000 older‑onset T2D patients [minor allele frequency (MAF)=0.007], 2/500 non‑diabetic controls (MAF=0.002) and 3 prediabetic individuals who were previously classified as non‑diabetic controls. In silico mutagenesis and protein modeling of p.H238N revealed changes of the polar contacts across the tetratricopeptide repeat (TPR) motif and TPR subdomains, which may affect the protein tertiary structure. Furthermore, the expression of DNAJC3 H238N protein was 0.68±0.08 fold (P<0.05) lower when compared with that of the wild‑type, possibly due to protein instability. Thus, DNAJC3 p.H238N is likely to be a variant causing diabetes.