NCDS

The 1958 National Child Development Study (NCDS) is following the lives of an initial 17,415 people born in England, Scotland and Wales in a single week of 1958. It started in 1958 at birth, as the Perinatal Mortality Survey. Cohort members were genotyped during the biomedical sweep (sweep 8, aged 44-45).

Data availability

Genetic QC

TOPMed QC

Genotyping for 13,738 samples (6,431 unique individuals) was performed across seven different genotyping arrays (see Genotyping arrays). Quality control was completed using PLINK1.9, PLINK2.0, R v3.3.2 and RStudio v4.1.2. Each dataset was updated to GRCh37 build for consistency using the LiftOver software tool. For each chip individuals were excluded if they had (i) they had > 2% missing data (ii) their genotype predicted sex using X chromosome homozygosity was discordant with their reported sex (excluding females with an F value > 0.2 and males with an F value < 0.8) (iii) they had excess heterozygosity [>3 standard deviation (SD) from the mean] and (iv) they were related to another individual in the sample (king-cutoff 0.0884)), removing samples with the most missing data. Prior to imputation SNPs with high levels of missing data (>3%), Hardy-Weinberg equilibrium P < 1x10-6 or minor allele frequency <1% were excluded.

The cleaned data were checked against the HRC reference panel r1.1 site list (Haplotype Reference Consortium, 2016) for strand issues, using a Perl script named HRC-1000G-check-bim.pl, v.4.3.0 (Rayner, Reference Rayner2020) from the Wrayner/ McCartney tools box. We used the HRC panel for this step since the data were genome build GRCh37 and were mostly European. The genetic data were then recoded as vcf files before uploading to the TOPMed which uses Eagle2 to phase haplotypes, and Minimac4 (https://genome.sph.umich.edu/wiki/Minimac4) with the TOPMed reference panel.

Imputed genotypes were then filtered with PLINK2.0alpha, excluding SNPs with an R2 INFO score < 0.8 and recoded as binary PLINK format. Samples with >2% missing values, and SNPs with >2 alleles (using –max-allele 2), >3% missing values, Hardy-Weinberg equilibrium P < 1e-6 or a minor allele frequency of <1% were excluded. We combined data from five of the seven chips (Illumina 1.2M, Illumina Human 660-Quad, Infinium HumanHap 550K v1.1, Infinium HumanHap 550K v3 and Affymetrix v6) which had, high and similar imputation quality (based on number of SNPs after QC). The Illumina 15k Custom Chip and Affymetrix 500k were not included in the combined dataset since they produced lower quality results, yielding less high-quality imputed SNPs than the other arrays. It is worth noting all samples, bar five, were covered by the other arrays and the combined dataset consisted of 6471 individuals and 7,545,768 SNPs. We have included a column within the basic demographics file indicating which samples are included from which chip in the final combined dataset.

Further QC was conducted on the combined dataset where individuals were excluded if they had (i) they had > 2% missing (0 individuals excluded) and (ii) they were related to another individual in the sample (king-cutoff 0.0884) (23 excluded), where one individual from each pair of related samples was excluded based on the King greedy related algorithm. Samples which could not be included in the sample (e.g. withdrawn consent) were removed (52 exluded). SNPs with high levels of missing data (>3%) (0 variants excluded) and a Hardy-Weinberg equilibrium P < 1e-6 were excluded (60 variants excluded).

We identified European samples using the GenoPred pipeline, which involves (i) merging the NCDS genotypes with data from 1000 genomes Phase 3, (ii) linkage disequilibrium pruning the overlapping single nucleotide polymorphisms (SNPs) such that no pair of SNPs within 1000 bp had r2 > 0.20 and (iii) using an elastic net model to establish which of the super populations the samples fall into (Africans [AFR], Admixed Americans [AMR], East Asians [EAS], Europeans [EUR] and South Asians [SAS]). Although each sample gets assigned to a superpopulation, there are some ancestral outliers within these groups. We used a threshold of >4 standard deviations from mean in PC1 to exclude outliers (72 excluded from GenoPred + PCA analyses). We have not removed these samples but a flag has been included in the basic dempgraphics file so that researchers can limit their samples to Europeans based on this. The final dataset consists of 6,396 samples and 7,496,556 genetic variants (genome build: hg38). The data are provided in plink binary format.

HRC QC

Genotyping for 13,738 samples (6,431 unique individuals) was performed across seven different genotyping arrays (see Genotyping arrays). Quality control was completed using PLINK1.9, PLINK2.0, R v3.3.2 and RStudio v4.1.2. Each dataset was updated to GRCh37 build for consistency, using up-to-date strand files and a series of commands collated in the script Update Build (Robertson, 2012) or using the LiftOver software tool. For each chip individuals were excluded if they had (i) they had > 2% missing data (ii) their genotype predicted sex using X chromosome homozygosity was discordant with their reported sex (excluding females with an F value > 0.2 and males with an F value < 0.8) (iii) they had excess heterozygosity [>3 standard deviation (SD) from the mean] and (iv) they were related to another individual in the sample (–genome threshold 0.1875), removing samples with the most missing data. Prior to imputation SNPs with high levels of missing data (>3%), Hardy-Weinberg equilibrium P < 1x10-6 or minor allele frequency <1% were excluded. The genetic data were then recoded as vcf files before uploading to the Michigan Imputation Server which uses Eagle2 to phase haplotypes, and Minimac4 (https://genome.sph.umich.edu/wiki/Minimac4) with the HRC r1.1 reference panel.

Imputed genotypes were then filtered with PLINK2.0alpha, excluding SNPs with an R2 INFO score < 0.8 and recoded as binary PLINK format. Samples with >2% missing values, and SNPs with >2 alleles, >3% missing values, Hardy-Weinberg equilibrium P < 1e-6 or a minor allele frequency of <1% were excluded. We combined data from five of the seven chips (Illumina 1.2M, Illumina Human 660-Quad, Infinium HumanHap 550K v1.1, Infinium HumanHap 550K v3 and Affymetrix v6) which had, high and similar imputation quality (based on number of SNPs after QC). The Illumina 15k Custom Chip and Affymetrix 500k were not included in the combined dataset since they produced lower quality results, yielding less high-quality imputed SNPs than the other arrays. It is worth noting all samples, bar five, were covered by the other arrays and the combined dataset consisted of 6420 individuals and 6,722,830 SNPs. We have included a column within the basic demographics file indicating which samples are included from which chip in the final combined dataset.

Further QC was conducted on the combined dataset where individuals were excluded if they had (i) they had > 2% missing (9 individuals excluded) and (ii) they were related to another individual in the sample (king-cutoff 0.0884) (16 excluded), where one individual from each pair of related samples was excluded based on the King greedy related algorithm. SNPs with high levels of missing data (>3%) (56,268 variants excluded) and a Hardy-Weinberg equilibrium P < 1e-6 were excluded. Samples were further excluded if they were classed as non-European, determined by merging the NCDS combined genotypes with data from 1000 genomes Phase 3), linkage disequilibrium pruning the overlapping single nucleotide polymorphisms (SNPs) such that no pair of SNPs within 50 bp had r2 > 0.20 and visually inspecting the first two genetic principal components along with the known ethnicities of the 1000 genomes sample to define European samples (N=83 excluded). The final quality controlled imputed set of genotypes contained 6312 samples and 6,663,631 variants (genome build: GRCh37/ hg19).

When using this dataset please cite: Bridges, E. C., Rayner, N. W., Mountford, H. S., Bates, T. C., & Luciano, M. (2023). Longitudinal Reading Measures and Genome Imputation in the National Child Development Study: Prospects for Future Reading Research. Twin Research and Human Genetics, 1-11.

The imputed data were QC’d and combined by CLS. QC and imputation of the Illumina Human 660-Quad was completed by CLS.

Genotyping arrays used in the HRC QC’d data set

Here is information on the individual genotyping chips used in the combined dataset. Breakdown of number of participants sampled on each array in the NCDS after removal of exclusions and duplications, and number of SNPs sequenced in each dataset (Bridges et al., 2023).

Other genetic data availability

All genotyping arrays (none-QCd)

If you would like to use the none-QCd data you can apply for the data from the separate genotyping chips. It is worth noting that these have been generated on different genotyping platforms, have multiple genome builds and some indiviuals have multiple samples.

DNA methylation data

*N=179 measured at both 44 + 62y

DNA methylation data preprocessing

DNA methylation data preprocessing was conducted at Bristol for batch 1 and at CLS for batch 2. Similar QC measures have been applied to both.

All data were QCd within R. Raw data for all datasets were used, prior to any quality control or normalization, and processed using the wateRmelon (Pidsley et al., 2013) package. Our stringent DNAm quality control pipeline included the following steps: i) excluding poorly performing samples based on methylated and unmethylated signal intensities; (ii) removing samples with a bisulphite conversion rate <80%; (iii) confirming reported sex through multidimensional scaling of X and Y chromosome sites; (iv) utilizing single nucleotide polymorphism (SNP) probes to verify genetic identity of matched samples; (v) employing the pfilter() function to exclude samples with >1% of probes with a detection P-value > 0.05; (vi) performing principal component analysis to identify and exclude outliers; and (vii) removing cross-hybridizing and SNP probes (Chen et al., 2013). Subsequent normalization of DNA methylation data was conducted using the nasen() function in wateRmelon (Pidsley et al., 2013). Any samples which have withdrawn consent or did not map to an ID have additioanlly been removed.

After QC, batch 1 consisted 529 individuals and 764,572 DNA methylation probes. Batch 2 consisted of 1298 samples (1119 individuals) and 804,070 DNA methylation probes.

Exome sequencing data