Fig 1.
Complex strategy for zygosity determination based on the three question of formal zygosity test, previously perceived zygosity and confused by strangers.
For example, if someone had “Yes” of Formal zygosity test, his zygosity would be the answer of question Previously perceived zygosity (PPZ) either “Monozygotic” or “Dizygotic” (but when he chose “Do not know” of question PPZ, his zygosity would be based on question Confused by strangers). On the contrary, when he chose “No” or “Do not know” in Formal zygosity test; his zygosity would be directly based on answers of Confused by strangers. In the question of Confused by strangers, “Yes” indicated to be identical twin, “No” referred to fraternal twin, “Hard to say” implied zygosity remained unclear.
Fig 2.
Relatedness Analysis of Discriminating Monozygotic\Dizygotic Twin Pairs by Genotyping and Methylation Array.
(a) was the plot based on Genotyping Array Data, (b) was the plot based on Methylation Array Data. Z0, Z1 referred to identity-by-descent (IBD) probabilities: those for the individuals having zero or one pairs of IBD alleles. The Blue Dots indicated Monozygotic Twins (MZ), which Both Z0 and Z1 Approximately equaled to 0, while the Red Dots indicated Dizygotic Twins (DZ), which Z0 was close to 0.25 and Z1 was close to 0.5 by Genotyping Array Data. (c) was an scatter plot of methylation 65 SNPs beta values in a representative DZ pair with calculating the intra pair correlation coefficient 0.7478, (d) was an example of MZ pair with correlation coefficient 0.9988, (e) was the boxplot of methylation 65 SNPs correlation coefficient for MZ versus DZ, the grey area indicated possible cut-off points (correlation coefficient ranged 0.84–0.90) for discriminating MZ and DZ twins.
Table 1.
Characteristics of the 192 same sex twin pairs who were analyzed for questionnaire based zygosity determination.
Table 2.
Logistic models among 192 same sex twin pairs of questionnaire based information in predicting DNA-determined zygosity.
Fig 3.
The discriminative ability compared among six logistic regression models by questionnaire zygosity determination.
(a-f) Boxplot of probabilities for Dizygotic Twins (DZ) from the Monozygotic Twins (MZ) versus DZ logistic regression Model 1–6, and (g) ROC curves showing the discriminative ability of the MZ versus DZ under logistic regression model. Model 1–6 were denoted by a red, yellow, orange, green, pink and blue line, respectively. The Areas Under the Curve (AUC) and its 95% Confidence Interval was 64.54% (57.32%-71.76%), 82.14% (76.34%-87.95%), 77.55% (71.47%-83.64%), 86.49% (81.07%-91.91%), 88.67% (83.50%-93.85%) and 89.03% (83.86%-94.19%) for Model 1–6.