Pre-processed GWAS summary statistics for Amyotrophic Lateral Sclerosis (ALS) from literature studies. This dataset has been cleaned and standardized for use in B parameter estimation (effect size as a function of MAF).
ALS_snvs_B_trainingA data frame with 75 rows (variants) and 9 columns:
SNP identifier (character)
Chromosome, autosomes only (numeric 1-22)
Genomic position in GRCh37 (numeric)
Minor allele frequency, 0 < MAF \(\le\) 0.5 (numeric)
P-value, 0 < P \(\le\) 1 (numeric)
Sample size, N \(\ge\) 1000 (numeric)
Effect size as log odds ratio (numeric)
"binary" for all variants (character)
Study accession identifier (character)
GWAS Catalog https://www.ebi.ac.uk/gwas/
Prepared from raw ALS literature data using prepare_B_training_data
with the following quality control:
Autosomal variants only (CHR 1-22)
Minimum sample size: N \(\ge\) 1000
Non-diagnostic traits excluded (age of onset, survival, C9orf72 interaction)
Nicolas A et al. (2018) study excluded (see note below)
Duplicates removed (kept variant with largest N)
Rows with NA in MAF, P, N, or BETA removed
The Nicolas A study is excluded because its variants overlap substantially with other included studies but report from a different analysis model. This validated 75-SNV dataset produces B estimates with R-squared = 0.50 using the beta method, and correlation > 0.9999 with the original reference formula.
Provenance: prepared from the curated 75-SNV ALS set; the raw input
inst/extdata/ALS-known-SNPs-raw.xlsx ships with the package.
# Load the data
data(ALS_snvs_B_training)
# Examine structure
str(ALS_snvs_B_training)
#> 'data.frame': 75 obs. of 9 variables:
#> $ rsID : chr "rs6703183" "rs16856202" "rs873917" "rs12145183" ...
#> $ CHR : num 1 1 1 1 10 10 10 11 11 11 ...
#> $ POS : num 2.10e+08 2.32e+08 4.01e+07 7.53e+07 1.14e+08 ...
#> $ MAF : num 0.3416 0.04 0.29 0.0762 0.348 ...
#> $ P : num 1e-08 8e-06 8e-06 2e-14 7e-10 ...
#> $ N : num 4654 4079 4079 4654 138086 ...
#> $ BETA : num 0.262 0.693 0.148 0.438 0.06 ...
#> $ TRAIT_TYPE: chr "binary" "binary" "binary" "binary" ...
#> $ STUDY : chr "GCST009529" "GCST000406" "GCST000406" "GCST009529" ...
#> - attr(*, "column_mappings")=List of 11
#> ..$ BETA : chr "Beta_numeric"
#> ..$ MAF : chr "RAF"
#> ..$ P : chr "P-value"
#> ..$ N : chr "Sample Size"
#> ..$ rsID : chr "rsID"
#> ..$ CHR : chr "chr"
#> ..$ POS : chr "pos_GRCh37"
#> ..$ STUDY : chr "Study accession"
#> ..$ PMID : chr "PubMed ID"
#> ..$ FIRST_AUTHOR: chr "First Author"
#> ..$ OR : chr "OR"
#> - attr(*, "mapping_sources")=List of 11
#> ..$ BETA : chr "user-specified"
#> ..$ MAF : chr "user-specified"
#> ..$ P : chr "user-specified"
#> ..$ N : chr "user-specified"
#> ..$ rsID : chr "user-specified"
#> ..$ CHR : chr "user-specified"
#> ..$ POS : chr "user-specified"
#> ..$ STUDY : chr "user-specified"
#> ..$ PMID : chr "user-specified"
#> ..$ FIRST_AUTHOR: chr "user-specified"
#> ..$ OR : chr "auto-detected"
head(ALS_snvs_B_training)
#> rsID CHR POS MAF P N BETA TRAIT_TYPE
#> 1 rs6703183 1 209712889 0.3416 1e-08 4654 0.26236426 binary
#> 2 rs16856202 1 232155151 0.0400 8e-06 4079 0.69314718 binary
#> 3 rs873917 1 40132795 0.2900 8e-06 4079 0.14842001 binary
#> 4 rs12145183 1 75272480 0.0762 2e-14 4654 0.43825493 binary
#> 5 rs58854276 10 114145044 0.3480 7e-10 138086 0.06000000 binary
#> 6 rs3736947 10 114169190 0.3500 8e-11 84343 0.08342161 binary
#> STUDY
#> 1 GCST009529
#> 2 GCST000406
#> 3 GCST000406
#> 4 GCST009529
#> 5 GCST90027163
#> 6 GCST010857
if (FALSE) { # \dontrun{
# Use in B estimation
B <- get_B(
training_trait = "binary",
training_MAF = ALS_snvs_B_training$MAF,
training_BETA = ALS_snvs_B_training$BETA,
target_trait = "binary",
target_MAF = seq(0.01, 0.4, by = 0.05),
target_case_prop = 0.5
)
} # }