Sibling variation in polygenic traits and DNA recombination mapping with UK Biobank and IVF family data

Abstract

We use UK Biobank and a unique IVF family dataset (including genotyped embryos) to investigate sibling variation in both phenotype and genotype. We compare phenotype (disease status, height, blood biomarkers) and genotype (polygenic scores, polygenic health index) distributions among siblings to those in the general population. As expected, the between-siblings standard deviation in polygenic scores is [Formula: see text] times smaller than in the general population, but variation is still significant. As previously demonstrated, this allows for substantial benefit from polygenic screening in IVF. Differences in sibling genotypes result from distinct recombination patterns in sexual reproduction. We develop a novel sibling-pair method for detection of recombination breaks via statistical discontinuities. The new method is used to construct a dataset of 1.44 million recombination events which may be useful in further study of meiosis.

Figure 1

Trait and genetic variation among siblings is smaller than in the general population, but still significant. Population genetics theory implies that the standard deviation in sibling PGS is $\sqrt{2}$ smaller than for the general population. We demonstrate this empirically using UKB and IVF clinic data, and illustrate that variation in important traits and disease risks among siblings is substantial. Note that the exact $\sqrt{2}$ relation only holds at the phenotypic level if the trait is 100% heritable.

Figure 2

Reduced, but still significant, variation among siblings relative to general population. UKB sibling pair vs unrelated pair differences in phenotype (vertical axis) and corresponding difference in PGS (horizontal axis). The left and right panels are height and BMI, respectively.

Figure 3

Type 2 Diabetes PGS: sibling variation smaller but still significant. Difference in PGS for Type 2 Diabetes in pairs of siblings and pairs of unrelated individuals from the general UKB population. The width of the sibling distribution is, as expected, $\sim \sqrt{2}$ times smaller, but still significant. The dashed lines mark the standard deviations for the distributions.

Figure 4

Health Index Polygenic Score: reduced, but still significant, variation among siblings relative to general population. For UKB sibling pair vs unrelated pair differences in Health Index we also find $\sqrt{2}$ reduction in standard deviations of PGS differences for siblings, as indicated by the dashed lines.

Figure 5

Type 2 Diabetes PGS by family (uncentered) illustrates significant variation within each family. Histogram on left describes UKB general population. Sets of siblings organized by family are displayed on the right to indicate both inter- and intra-family variation as compared to the general population. Note in some families all siblings are above or below the general population average in PGS. Horizontal lines indicate the standard deviation of the general UKB population.

Figure 6

Type 2 Diabetes Polygenic Risk Score for UKB families (centered). Histogram on far left describes UKB general population. Middle histogram describes sibling scores after recentering relative to sibling average score (parental midpoint is generally not available). Sets of siblings organized by family are displayed on the right to indicate intra-family variation as compared to the general population. The sibling values have been centered within each family, unlike in the previous figure.

Figure 7

Type 2 Diabetes Polygenic Risk Score for GP/IVF families (uncentered). Far left histogram describes UKB general population. Middle histogram describes GP IVF embryos in aggregate (note this is a mixture of highly-related (sibs) (sibs from different families). Individual families are displayed on the right with mother represented by a circle and father by a square. Note siblings can have higher or lower PGS than either parent. Horizontal lines indicate standard deviations in the general populations.

Figure 8

Type 2 Diabetes Polygenic Risk Score for GP/IVF families (centered). Far left histogram describes GP embryos in aggregate. Middle histogram describes GP IVF embryos after recentering with respect to parental midpoint. Individual families are displayed on the right with mother represented by a circle and father by a square. Families on right are displayed after recentering at midparent PGS. Again, the population standard deviations are marked with horizontal lines.

Figure 9

Genetic Health Index value distributions in the UK Biobank are similar to those of GP parents and embryos. Top-left: UK Biobank vs all GP samples. Top-right: GP parent samples vs GP embryo samples. Bottom-left UK Biobank vs parent samples. Bottom-right: UK Biobank vs embryo samples.

Figure 10

Genetic Health Index for GP/IVF families (uncentered). Far left histogram describes UKB general population. Middle histogram describes GP IVF embryos in aggregate (note this is a mixture of highly-related (sibs) (sibs from different families). Individual families are displayed on the right with mother represented by a circle and father by a square. Note siblings can have higher or lower Index than either parent.

Figure 11

Genetic Health Index for GP/IVF families (centered). Far left histogram describes GP embryos in aggregate. Middle histogram describes GP IVF embryos after recentering with respect to parental midpoint. Individual families are displayed on the right with mother represented by a circle and father by a square. Families on right are displayed after recentering at midparent Index.

Figure 12

Cumulative recombination length of chromosomes 1 and 22 (i.e., total Morgan length). Red line is the siblings-pair method applied to 21 k European ancestry sibs from UK Biobank. Gray line is from deCODE (Science 2019) result using population of Iceland.

Figure 13

Density of recombination breaks on chromosomes 1 and 22 from sibling-pair method, computed using 21 k European sibling pairs.

Figure 14

The recombination algorithm detects breaks between regions in which both siblings have identical SNPs and regions where they differ. SNP genotypes for sibling 1 and 2 are shown as counts of the major alleles. The recombination break occurs at the boundary between regions. On the left both siblings have inherited the same chromatid strands from the parents. A recombination break, which could occur in either the egg or the sperm progenitor of either sibling, causes mismatches in the SNP genotypes in the region on the right—i.e., on the right the siblings no longer have the same chromatid strands.