Biomedical subjects
Publications and source records attributed to J C Christian.
We discuss informative plots for univariate twin data that can be used in conjunction with twin data analyses. The plots are useful for spotting outliers, spotting possible single gene effects, and displaying the contribution of individual twin pairs to the fit of genetic models of the data. We illustrate the use of the plots on bone mineral data, and present programs for generating the plots in SAS.
Explore the source record for details and available documents.
We discuss the program, TWINAN90, which can perform several different types of analysis of twin data. TWINAN90 incorporates the ANOVA-based twin analyses from the TWINAN twin analysis program, and also includes maximum likelihood estimation of parameters from three path models. Another feature of TWINAN90 is the optional output of a pedigree file which can be read by the quantitative genetics package FISHER. The diagnostic features of the program make TWINAN90 useful also for preliminary analyses prior to the use of more sophisticated modeling procedures which are available in packages such as LISREL and FISHER. An annotated printout from TWINAN90 is presented to illustrate the statistical analyses performed in the program.
The multiple relationships within kinships of adult monozygotic twins permit incisive analyses to be made of genetic and environmental effects on behavioral traits. Data from families of 65 monozygotic twin pairs yield evidence of genetic variance on the Block Design Test, a nonverbal measure of general intelligence.
To search for an association of twin zygosity with tooth size, 56 dental variables measured from 65 pairs of twins (43 MZ, 22 DZ) were studied. Results of the t' test for equality of the means showed no association of zygosity with any of the variables in males or in females. Results of the F' test for homogeneity of total variances between zygosities showed evidence for unequal total variances in 15 variables in males and 13 in females. Sex influence was further noted on the association of zygosity with the variance of tooth size. Where total variances were unequal, genetic variance estimates differed when only the within-pair mean squares were used and when combined estimates designed to be unbiased by differences in environmental variances were used.
Possible genetic influences on glomerular filtration rate and electrolyte excretion were investigated in 55 (37 monozygotic, 18 dizygotic) young adult white twin pairs. Subjects were studied during a five-day hospitalization involving sodium loading and sodium depletion. No evidence of genetic variability was found in the control levels of serum or urine sodium and potassium. Following a saline infusion it was possible to detect genetic influence in electrolyte handling. Creatinine clearance, used as a measure of glomerular filtration rate, did not appear to be genetically mediated. The results indicate that genetic factors are important in sodium handling in normal individuals and that this is independent of glomerular filtration rate.
Analysis of variance was performed on the radial and ulnar finger ridge counts and ridge count diversity index in 360 twin sets from which estimates of genetic variance were obtained. Findings for radial and ulnar counts paralleled those previously obtained for finger pattern type and ridge count (larger of radial and ulnar count). In contrast, ridge count diversity showed no indication of unequal total variances, as previously found for the total ridge count. One must be cautious not to exclude genetic influences on traits, such as the thumb variables in this study, where there are unequal total variances between monozygotic and dizygotic twins and the more conservative estimate of genetic variance is not significant.
Twin methodology for the analysis of continuously distributed traits is presented as a guideline for initial steps in the analysis of twin data. Tests for heterogeneity of twin means and variances are reviewed along with estimates of genetic variance. Analysis of twin data appears on the surface to be a rather simple task, but investigators may well be confused by the wealth of choices available for hypothesis testing. This paper is designed to help the less experienced investigator, with perhaps a modest amount of data, begin analysis with maximum attention to testing all testable assumptions of the twin model, and using recent examples of applications of the methods.
Birthweight was measured on 188 monochorionic monozygotic, 54 dichorionic monozygotic, 102 like-sexed dizygotic, and 94 unlike-sexed dizygotic liveborn twin pairs. Overall, males were found to be significantly heavier than females. These differences were not significant, however, when birthweights were compared within zygosity/chorion-type categories. Males were also characterized by a slightly greater overall total variance. Comparisons of intrapair variation of monochorionic and dichorionic monozygotic twins revealed significant differences between monochorionic pairs and dichorionic separate pairs and no significant differences between monochorionic pairs and dichorionic fused pairs. The results of this study suggest that placental proximity may have as important an influence on variation in birthweight as does the presence or absence of vascular anastomoses.
In a random sample of 104 pairs of middle-class Caucasian, handedness-discordant twins six years of age or older, a significant relationship has been found between birth order and handedness in monozygotic twins, there being an excess of left-handed among first-born twins (P less than 0.01). No such relation has been found in dizygotic twins.
The multiple relationships which exist within families of adult monozygotic (MZ) twins permit evaluation of genetic and environmental variance and provide a direct test of maternal effects. Systolic blood pressure (SBP) was measured in 610 members of the families of 76 MZ twin pairs. Age and sex effects on BP were removed via a Z-transformation. Correlations for SBP were .72 for MZ twins and .23 for full siblings. Regression of offspring on single parent was .26. The husband-wife correlation was significant, but it was no higher than the resemblance of a twin to the spouse of his co-twin. Comparisons of maternal and paternal half-siblings provide no evidence of maternal influences on SBP. To evaluate whether familial resemblance of BP is mediated by body size, the original systolic pressures were adjusted by stepwise multiple regression of age, sex, height, weight and skinfold thickness. Adjusting BP for effects of body size reduced regression of offspring on mid-parent from .40 to .23. The authors conclude that familial aggregation of systolic pressure reflects additive genetic variance mediated, in part, by body size and augmented by social homogamy arising from non-random mating.
Significantly larger variation between sibships within families of male MZ twins than between sibships within families of female MZ twins, indicative of maternal influences, was found for 10 of 41 dermatoglyphic fingertip variables. Of these, five were thumb-related with the effect primarily on the thumb radial and ridge count (larger of radial and ulnar count). These same variables were previously found to have unequal variances in MZ twins of known placental type, and the results indicate maternal influences in singletons as well as twins for these variables. Although the total ridge count (TRC), previously shown to differ in MZ twins of known placental type (paralleling the thumb radial and ridge counts) did not reach significance, the trend indicated that the observed thumb changes may be reflected in the TRC as well. Little finger pattern type and ulnar counts also showed less variability in families of female MZ twins, but the interpretation is complicated by the concomitant differences in mean squares within-sibships for these little finger variables.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Over the past 50 years a large number of methods have been proposed for estimating heritability from twin studies. The present paper describes the most commonly cited of these estimates as a first step in evaluating their usefulness. A critical review will then follow.
Factor analysis was employed using the ulnar ridge count, radial ridge count, ridge count (the larger of the radial or ulnar count as generally used for calculating total ridge count), and pattern type for each finger in 720 twins. Pattern type and ulnar count displayed parallel factor loadings while loadings for radial and ridge count also paralleled each other. This relationship did not hold for the index finger, indicating the importance of pattern direction and greater pattern diversity for this digit. Total ridge count was most closely associated with a factor of ring and little finger radial and ridge count and only secondarily with an index finger factor. When radial and ulnar counts were deleted to make comparisons with earlier studies, the result was factors having groupings of variables identical with previous reports. It appears that factor analysis results in consistent extraction of identical or very similar factors from different populations, and the use of radial and ulnar counts adds more information than when only the larger of the two counts is considered.
Explore the source record for details and available documents.