1. Background:

A previous paper involving the two cohorts, Sorbs and German Obesity Trios, identified alleles at some SNPs that differentially impact offspring depending on whether they are transmitted from the father or the mother. We use the same method to replicate the key findings, but with the Framingham cohort. The key findings from the different cohorts are summarized as follows:

1.1 Sorbs Cohort:

1.1.1 Three SNPs from Intron_1 (rs1861868, rs1121980, and rs9939973) show POE(Parent-of-Origin).

1.1.2 rs1121980, and rs9939973 show significant positive and stronger paternal effect size to offspring BMI.

1.1.3 rs1121980 and rs9939973 are in high LD (Linkage Disequilibrium) with rs8050136, one of the well replicated variants strongly associated with BMI.

1.2 German Obesity Trios Cohort:

1.2.1 One SNP (rs10852522) from Intron_2, two SNPs (rs17818920, rs17818902) from Intron_3 are observed with POE, and maternal alleles are showing higher transmission rates.

1.3 Previous Results in Framingham Cohort:

1.3.1 Our previous study using the Framingham cohort focused on single SNP analysis and indicated that two SNPs (rs1121980, rs9939973) located in intron_1 show a trend (p ≈ 0.1) toward parent-of-origin effects (POE). Both paternal and maternal alleles affect in the same direction, although the maternal allele has a stronger effect size

1.3.2 In contrast, three SNPs (rs7203051, rs8053740, rs7205009) from intron_3 show significant POE (p<0.05), where the effects of paternal and maternal alleles are in opposite directions, with a stronger effect size observed in the paternal allele.

2. Hypothsis:

We assume that additional SNPs from Intron_1 to Intron_3 will show POE on offspring BMI, and the POE is the same across the intron regardless of the SNPs, the maternal alleles always show stronger effect, or vice versa.

3. Methods:

We would like to expand the search range to include all SNPs from Intron 1 to Intron 3 of the FTO gene on Chromosome 16, with the goal of identifying additional SNPs that may indicate POE.

Step 1: Use the Genome Browser with the human assembly from February 2009 (GRCh37/hg19) to extract all SNPs from Intron 1 to Intron 3 of the FTO gene on Chromosome 16.

Step 2: Filter the extracted SNPs by retaining only those with a Minor Allele Frequency (MAF) greater than 0.01 and an imputation quality (R²) greater than 0.8.

Step 3: Conduct a Linkage Disequilibrium (LD) analysis to identify SNP pairs that are highly correlated, with an R² greater than 0.8.

Step 4: Group highly correlated SNPs together and, within each group, order the SNPs according to their genomic coordinates.

Step 5: For each group, concatenate the first allele of each genotype as the paternal haplotype and the second allele as the maternal haplotype. Perform haplotype analysis separately for the paternal and maternal haplotypes. Filter the results to retain only the groups that show significance in either the paternal or maternal haplotypes alone, or in both.

Step 6: To simplify contrast analysis computations, the SNP with the highest MAF was selected from each group to directly compare paternal and maternal inheritance effects.

Step 7: Use a linear regression model to contrast the effects of paternal and maternal inheritance by analyzing the heterozygous genotypes.

4. Results:

Intron_1: Three SNPs (rs74020804, rs12447107, rs28711250) exhibit POE. Paternal inheritance shows a significant effect, while maternal inheritance does not. The effects are in opposite directions, with the paternal association having a negative but stronger effect size compared to the maternal association.

Intron_2: One SNP(rs75547181) demonstrates POE. Paternal inheritance shows a significant effect, while maternal inheritance does not. The effects are in opposite directions, with the paternal association having a negative but stronger effect size compared to the maternal association.

Intron_3: Two SNPs (rs16952570, rs2388407) show POE. For SNP rs16952570, the maternal effect is significant, while the paternal effect is not. Conversely, for SNP rs2388407, the paternal effect is significant, while the maternal effect is not. Both SNPs have a negative paternal impact and a positive maternal impact on offspring BMI. However, rs16952570 has a stronger positive maternal effect, whereas rs2388407 exhibits a stronger negative paternal effect.

5. Conclusion

Therefore, we partially support the hypothesis, as among the six SNPs of POE from Intron 1 to Intron 3, five SNPs exhibit a stronger negative paternal effect size, while one POE SNP from Intron 3 demonstrates a stronger maternal effect size.

6. Discussion

The table below combines the results of haplotype analysis and single SNP analysis from the Framingham cohort, along with results from the Sorbs cohort and the German Trios cohort. The green-highlighted cells represent POE SNPs along with their corresponding p-values from the Framingham haplotype analysis. The grey column indicates single SNPs of interest mapped to their associated haplotype groups. Pink cells display the p-values of single SNPs showing POE from the Framingham cohort, while yellow cells show the p-values of POE SNPs from the Sorbs cohort, and blue cells represent the p-values of POE SNPs from the German Obesity Trios.

Correlated multiple SNPs won’t affect offspring BMI differently compared to single SNPs. The single SNP and the haplotype group containing multiple SNPs demonstrate the same parental inheritance pattern when affecting the offspring BMI. For example, the three SNPs(rs7203051, rs8053740, rs7205009) from Intron 3 that show POE from single SNP analysis from Framingham cohort are classified within haplotype group 5, which consists of 27 SNPs. Both the single SNP and group 5 exhibit POE, with paternal and maternal effects in opposite directions, where the paternal association has a negative but stronger effect size compared to the maternal association.

None of the Framingham, Sorbs and German Trios shares the same POE SNP.

Recap:

Sorbs: Intron_1 (rs1861868, rs1121980, and rs9939973) show POE. rs1121980, and rs9939973 show significant positive and stronger paternal effect size to offspring BMI.

German Trios: Intron_2 (rs10852522), Intron_3 (rs17818920, rs17818902) show POE, and maternal alleles are showing higher transmission rates.

Framingham Single SNP Analysis:
Intron_1 (rs1121980, rs9939973) exhibits a trend toward POE, both paternal and maternal alleles show effects in the same direction, maternal is stronger. Intron_3 (rs7203051, rs8053740, rs7205009) shows POE, where the effects of paternal and maternal alleles are in opposite directions, paternal is stronger.

Framingham Haplotype Group Analysis: Intron_1 (rs74020804, rs12447107, rs28711250), Intron_2 (rs75547181), Intron_3 (rs16952570, rs2388407) show POE. Paternal and maternal effects are in the opposite direction. Only rs16952570 from Intron 3 exhibits a stronger maternal effect, while all other five SNPs demonstrate stronger paternal effect on offspring BMI.

Note: You may find duplicate FHS_Hap rows in the table because the group contains more than one single SNP of interest. Expanding the group allows for better comparison with the single SNP analysis.