Numerous studies have shown that longevity is moderately heritable in human populations. Longevity, however, contains limited information on functional status, since individuals may exhibit differential survival patterns. In this study, we employed a stepwise multiple regression approach to estimate biological aging in a Mennonite population, using chronological age as a dependent variable and various predictors of chronological age including subphenotypes related to diabetes, coronary heart disease, hypertension, renal function, and markers of functional ability. The residual (the difference between chronological and predicted ages) is considered a marker of biological age. In fact, two different data sets were used to obtain residuals due to the availability of data. In each analysis, chronological age was regressed on predictor variables in a stepwise manner, retaining the variables significant at the 5% level. The first analysis (N=729) included 6 significant predictors (R(2)=44.3%): glucose, blood urea nitrogen (BUN), cholesterol, albumin, systolic blood pressure (SBP), and ln potassium, and the second analysis (N=232) included 9 significant predictors (R(2)=71.5%): BUN, albumin, SBP, low-density lipoprotein cholesterol, forced expiratory volume in 1 sec (FEV1), grip strength, trunk flexibility, reaction time, and FEV1xsex. Using a variance components approach, we found that the data set-specific residuals were significantly heritable (h(2)+/-SE): first analysis=0.265+/-0.106, and second analysis=0.469+/-0.180. The residuals from the second data set appear to be more informative for biological aging, perhaps due to the inclusion of functional ability-related phenotypes in addition to the blood chemistry variables. In summary, we have shown that markers of biological aging in Mennonites are under substantial additive genetic influences.