Dermatoglyphics of schizophrenics patients with Down's syndrome and mentally retarded males as compared with Australian Europeans using multivariate statistics.код для вставкиСкачать
Dermatoglyphics of Schizophrenics, Patients with Down’s Syndrome and Mentally Retarded Males as Compared with Australian Europeans using MuIti var iate Statistics SARDOOL SINGH School of H u m a n Genetics, T h e University of New South W a l e s , Kensington N.S.W. 2033 KEY WORDS Dermatoglyphics . Multivariate analysis . Schizophrenia . Down’s syndrome . Mental retardation. ABSTRACT Dermatoglyphics of schizophrenics, patients with Down’s syndrome and mentally retarded males were compared with those of normal Australian Europeans. A computer programme of multivariate analysis of fifteen dermatoglyphic features was utilized. This analysis produces two significant variates of each of the populations plotted in two-dimensional space. The distance, measured in arbitrary units, between any two populations was studied for its significance. It was noticed that the patients with Down’s syndrome sepaarated significantly from the rest of the groups. The importance of multivariate analysis in the study of dermatoglyphics in comparing two or more populations is discussed. The use of multivariate analysis becomes necessary when several measurements are to be utilized in distinguishing two or more populations. Dermatoglyphic data are of this type because ten or more variables are always available. One or two discriminant functions, plotted as graphs, can show almost all available information about the differences between any two sampled groups by condensing information relating to a large number of variables. In this paper a multivariate analysis based on Anderson (’58) is used to study the divergence between normal Australian Europeans, schizophrenic patients, patients with Down’s syndrome and mentally retarded males. MATERIAL AND METHOD All populations studied were of European origin. A random sample of 500 males and 500 females was collected from the districts of N.S.W. The quantitative description of this sample has been presented by Singh (’67a). Eighty-nine male and 83 female schizophrenics were collected from Callan Park Mental HosFital, AM. J. PHYs. ANTHROP.,42: 237-240. Sydney. These patients were certified long term schizophrenics and have been hospitalized in “long stay wards” for upwards of six months. Their ages varied from 20 to 60 years. They have been quantitatively described in Singh (’67b). Ninety-four male and 62 female patients with Down’s syndrome, collected from various institutions for handicapped children in N.S.W., have been quantitatively reported in Singh (’68) . In addition a sample of 53 mentally retarded male patients was collected from Peat and Milson Island Hospital in N.S.W. This group does not include any mental disorders due to known genetic causes viz. Down’s syndrome or phenylketonuria. The cause of their mental deficiency was undiagnosed. It was of interest to study their dermatoglyphics because they were suffering from the same degree of mental dificiency as the patients with Down’s syndrome and in one study were used as controls for Down’s syndrome (Singh, ’69). Fifteen dermatoglyphic features of each population are used: the ridge counts on 1 Present address: Kanematsu Memorial Institute Sydney Hospital, MacQuarie Street, Sydney, N.S.W: 2000. 237 238 SARDOOL SINGH ten fingers, the frequencies of ulnar loops, radial loops and accidentals, and the triradius counts, and in addition the asymmetry, as defined by Singh ('68). The data of the above populations were available on punch cards. A multivariate analysis based on Anderson ('58) was used in the programme run on the CDC 3600 computer at the C.S.I.R.O. Division of Computing Research, Canberra. TABLE 2 Showing the linear combinatons o f observed variates comprising significant discriminant f u n c t i o n s f o r t h e comparison of all groups' RI RII RIII RIV RV LI LII LIII LIV LV Ulnar loops Radial loops Accidentals Triradius counts Asymmetry RESULTS Table 1 gives the canonical correlations and their significance with chi-square values and degrees of freedom. It will be seen that the first and second correlations are significant while the third and fourth are not. Combinations of observed canonical variates comprising significant discriminant functions are shown in table 2. From table 2 are calculated the co-ordinates (vectors), V, and V,, for plotting the mean positions of each population - these values are shown in table 3 ; the dispersion of population centroids is shown in figure 1. The separation of all males from females and of the patients with Down's syndrome from the rest of the populations can be seen. In this figure the distance measured between any two populations represents the generalized Mahalanobis D' statistic, the significance of which can be tested with a n F-ratio test calculated by the following formula : Coefficients of discriminant functions Variates v1 vz -0.11696 -0,14634 -0.11057 0.25383 0.04087 -0,11482 - 0.07693 - 0.38045 -0,30537 0.25669 -0.32012 0.15410 0.00703 0.08008 - 0.05632 -0,22435 -0.03127 0.03159 0.22703 - 0.22943 -0.31101 0.04705 - 0.07843 0.02413 -0,35300 -0.07575 0.08821 - 0.02311 0.28660 0.3368 1 The within group means and standard deviations of the populations used are taken from Singh ('67a, '67b, '68, '69). TABLE 3 Canonical variates f o r group centroids ___.__ __- Males - Females ~~ - Australian European Schizophrenic Down's Syndrome Mentally retarded males v1 vz v1 ~ vz -0.160 - 1.245 -0.285 - 1.08 - 0.344 - 1.41 - 0.252 - 0.932 -0.984 -0.981 - 1.094 - 1.537 -0.228 -1.356 - - V P values have been scaled by the ratio of canonical correlations after normalising the coefficients within each vector. Latent roots removed Correlation xz D.F. 1 P 0 1 2 3 0.41332 0.23691 0.14606 0.12653 401.53 144.33 65.08 35.49 90 70 52 36 0.0001 0.001 N.S. N.S. + with ( p l ) , ( N , Nz p - 1) degrees of freedom, where N, is the number in the irhpopulation and p is the number of characters used. The results of this analysis are recorded in table 4. It will be seen from this table that normal Australian and schizophrenic males differ significantly from patients with Down's syndrome. Normal Australian females are significantly different from female patients with Down's syndrome, but not from schizophrenic females. - TABLE 1 Canonical correlations, w i t h Bartlctt's chi-squared, degrees o f f r e e d o m and their significance, and numbers of latent roots removed, f o r Australian Europeans, schizophrenics, patients w i t h Dozun's syndrome and mentally retarded males ~ DISCUSSION The degrees of freedom i n the above table are given by the expression ( p - r ) x (q - r ) , where r is the number of roots removed, p i s the number of group variables and q is the number of measured variables. 1 Comparison of the dermatoglyphics of abnormal populations with the controls 239 COMPARATIVE DERMATOGLYPHICS -200 -1.50 “2 - 0.50 11 Ment.Retd.M. 0 *Schf.M. M* ODowns Aust. M. 1 I I 0 -0.50 I I -1.00 -1.50 I -200 “1 Fig. 1 Scatter diagram of males ( M ) and females ( F ) of schizophrenics (Schf.), patients with Down’s Syndrome (Downs), mentally retarded males (Ment. Retd. M.) and Australian Europeans (Aust.). using univariate statistics does reveal significant differences, although the differences are not conclusive. In a n earlier study (Singh, ’67) schizophrenic males showed significantly less arches and more accidentals while the schizophrenic females showed significantly more whorls and less ulnar loops as compared to their respective controls. An interesting feature of this analysis was the significantly high incidence of accidentals on the third digit of the left hand of the schizophrenic males. Mentally retarded males when compared with patients with Down’s syndrome showed significantly more whorls and radial loops and less ulnar loops. The mean total ridge count of the mentally retarded males was also found to be significantly greater from that of the Down’s syndrome (Singh, ’69). The frequency of ulnar loops in male Down’s syndrome patients was significantly greater on all fingers, with the exception of fifth finger on both hands, as compared to normal males. In spite of the observed differences in some parameters between the abnormal populations and their controls, it cannot be said with any certainty that these differences are biologically significant. In a situation such as this where one is confronted with many variates, about 5% of them will show significant differences merely by chance. Multivariate statistics, however have a n advantage over the univariate statistics in that many varieties in many populations can be utilized simultaneously and the resultant figures tested for their significance. In a preliminary analysis it was found that whorls and arches contributed very little to the discriminant functions and hence have been omitted in the present analysis. The contribution to the betweengroup variance made by a variate is proportional to the squares of the coefficients of that variable in V, and V2. It will be seen from table 2 that the main contribution to TABLE 4 D? v a l u e s , F ratios a n d their significance ~ ~~ - .- ~~~ D2 F Ni NB Significance Between males Australian and schizophrenic Australian and mentally retarded Australian and Down’s Syndrome Schizophrenic and mentally retarded Schizophrenic and Down’s Syndrome Mentally retarded and Down’s Syndrome 0.07 0.03 0.81 0.01 0.61 0.77 0.35 0.09 4.17 0.02 1.73 1.57 14 14 14 14 14 14 575 537 578 128 169 131 N.S. N.S. p < 0.1% N.S. P 5% N.S. Between female s Australian and schizophrenic Australian and Down’s Syndrome Schizophrenic and Down’s Syndrome 0.01 0.49 0.49 0.05 1.76 1.04 14 14 14 567 545 129 B e t w e e n sexes Australian Schizophrenic Down’s Syndrome 0.08 0.24 0.31 1.31 0.64 0.70 14 14 14 984 158 140 - ~~~ ~ ~~~~~ ~~~~~~~ ~~ ~ ~~ ~~ < N.S. p < 5% N.S. N.S. N.S. N.S. 240 SARDOOL SINGH V, comes from the ridge counts of the first four fingers of the right hand and the first, third, fourth and fifth fingers of the left hand, and ulnar and radial loops. V, is accounted for mainly by the right first, fourth and fifth, the left first and fifth fingers, and from total triradius counts, the main contribution coming from the left first and fifth fingers. In figure 1 it can be seen that the values of V2 for the groups under study separate males from females, whereas the values of V, show a good separation of patients with Down’s syndrome from the rest of the groups. It is worth noting that Down’s syndrome, which is known to be the expression of a chromosomal abnormality, gives rise to a dermatoglyphic difference which is very evident in this study and that mental retardation and schizophrenia, for which no strong evidence of the causative influence of genetic factors has yet been found, do not differ significantly in dermatoglyphic discriminant measure from the normal Australian European groups. ACKNOWLEDGMENTS The author wishes to thank Professor R. J. Walsh, Professor of Human Genetics, University of New South Wales, for helpful suggestions in writing this paper, Dr. P. J. Claringbold and Mr. N. H. Westwood, C.S.I.R.O. Divisions of Animal Genetics and Computing Research, for statistical advice and computer facilities. LITERATURE CITED Anderson, T. W. 1958 A n Introduction to Multivariate Statistical Analysis. Wiley, New York, Chapter 12. Rao, C. R. 1952 Advanced Statistical Methods in Biometric Research. Wiley, New York. Singh, S. 1967a Quantitative analysis of finger ridge counts in Australians of European ancestry. H u m a n Biol., 39: 368-373. 1967b Dermatoglyphics in schizophrenia. Acta Genet., Basel, 17: 348-356. 1968 A measure of asymmetry of finger ridge counts. Acta Genet., Basel, 18: 599-605. 1969 Ph.D. Thesis, University of N.S.W., Kensington, N.S.W., Australia.