Pre-processed GWAS summary statistics for Bone Mineral Density (BMD) from large-scale literature studies. This dataset has been cleaned and standardized for use in B parameter estimation (effect size as a function of MAF).
BMD_snvs_B_trainingA data frame with 1790 rows (variants) and 10 columns:
SNP identifier (character)
Chromosome (numeric)
Genomic position (numeric)
Minor allele frequency, 0 < MAF \(\le\) 0.5 (numeric)
P-value, 0 < P \(\le\) 1 (numeric)
Negative log10 p-value, i.e., -log10(P) (numeric)
Sample size, N \(\ge\) 1000 (numeric)
Effect size (numeric, may contain NA)
"continuous" for all variants (character)
Study identifier (character)
GWAS Catalog https://www.ebi.ac.uk/gwas/
Prepared from raw BMD/osteoporosis literature data (CSV from GWAS Catalog)
using prepare_B_training_data with the following quality control:
Three high-quality studies included: "An atlas of genetic influences on osteoporosis" (Morris 2019), "Identification of 153 new loci", "Identification of 613 new loci"
Binary trait "bone fracture" excluded
Minimum sample size: N \(\ge\) 1000
Rows with NA in MAF, P, or N removed
Duplicates NOT removed (retains variants from different studies)
This is a continuous trait dataset. For cross-trait prediction (e.g., BMD to
osteoporosis risk), use the p-value method in get_B.
Provenance: prepared from the curated BMD/osteoporosis literature summary statistics (raw inputs are maintained with the source study, not bundled).
Morris JA, Kemp JP, Youlten SE, et al. (2019). An atlas of genetic influences on osteoporosis in humans and mice. Nature Genetics, 51(2), 258-266. doi:10.1038/s41588-018-0302-x
# Load the data
data(BMD_snvs_B_training)
# Examine structure
str(BMD_snvs_B_training)
#> 'data.frame': 1790 obs. of 10 variables:
#> $ rsID : chr "1:114494567_CT_C" "1:220189136_TA_T" "1:85296916_TG_T" "1:8657844_AT_A" ...
#> $ CHR : int 1 1 1 1 1 10 10 10 10 10 ...
#> $ POS : int NA NA NA NA NA NA NA NA NA NA ...
#> $ MAF : num 0.374 0.178 0.115 0.166 0.227 ...
#> $ P : num 2e-11 3e-36 5e-09 1e-23 2e-12 ...
#> $ P_mlog10 : num 10.7 35.5 8.3 23 11.7 ...
#> $ N : num 394929 394929 394929 394929 394929 ...
#> $ BETA : num 0.0148 0.0338 -0.0183 0.0276 -0.0168 ...
#> $ TRAIT_TYPE: chr "continuous" "continuous" "continuous" "continuous" ...
#> $ STUDY : chr "Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mi"| __truncated__ "Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mi"| __truncated__ "Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mi"| __truncated__ "Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mi"| __truncated__ ...
head(BMD_snvs_B_training)
#> rsID CHR POS MAF P P_mlog10 N BETA
#> 1 1:114494567_CT_C 1 NA 0.373892 2e-11 10.69897 394929 0.0147978
#> 2 1:220189136_TA_T 1 NA 0.178305 3e-36 35.52288 394929 0.0337965
#> 4 1:85296916_TG_T 1 NA 0.114804 5e-09 8.30103 394929 -0.0182615
#> 5 1:8657844_AT_A 1 NA 0.165803 1e-23 23.00000 394929 0.0275657
#> 6 1:98403226_TTTGTTG_T 1 NA 0.226728 2e-12 11.69897 394929 -0.0167524
#> 7 10:123694303_AGTTATAAAC_A 10 NA 0.086143 2e-18 17.69897 394929 0.0316996
#> TRAIT_TYPE
#> 1 continuous
#> 2 continuous
#> 4 continuous
#> 5 continuous
#> 6 continuous
#> 7 continuous
#> STUDY
#> 1 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
#> 2 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
#> 4 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
#> 5 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
#> 6 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
#> 7 Identification of 613 new loci associated with heel bone mineral density and a polygenic risk score for bone mineral density, osteoporosis and fracture.
if (FALSE) { # \dontrun{
# Use in B estimation (continuous trait, p-value method)
B <- get_B(
training_trait = "continuous",
training_MAF = BMD_snvs_B_training$MAF,
training_P = BMD_snvs_B_training$P,
training_N = BMD_snvs_B_training$N,
target_trait = "binary", # Cross-trait: BMD -> osteoporosis risk
target_MAF = seq(0.01, 0.4, by = 0.05),
target_case_prop = 0.20
)
} # }