remove_snps.Rd
Removes CpG sites from the object if they overlap with common SNPs
remove_snps( m, populations = NULL, maf_threshold = 0.01, reduce_filtering = FALSE, forced = FALSE, keep = FALSE, n_chunks = 1, n_cores = 1 )
m |
|
---|---|
populations | Populations to use. Default is all. |
maf_threshold | The frequency threshold, above which the SNPs will be removed. Default is 0.01 |
reduce_filtering | If TRUE, the SNPs with a MAF < 0.1 will be evaluated and only the highly variable ones will be removed. Default FALSE. |
forced | the reduce_filtering is not recommended with less than 10 samples, but can be forced. Default is FALSE. |
keep | Do you want to keep the sites that were filtered out? In this case, the function will return with a list of wo methrix objects. |
n_chunks | Number of chunks to split the |
n_cores | Number of parallel instances. Can only be used if input data is in HDF5 format. |
methrix object or a list of methrix objects
Takes methrix
object and removes common SNPs. SNPs overlapping with a CpG site and have a minor allele frequency (MAF)
above a threshold in any of the populations used will be selected and the corresponding CpG sites will be removed from the methrix
object.
With the reduce_filtering option, SNPs with MAP < 0.1 will be further evaluated. If they show low variance in the dataset, there is probably no genotype
variability in the population, therefore the corresponding CpG site won't be removed. Please keep in mind that variance thresholds are
#>#>#> chr N #> 1: chr21 42 #> 2: chr22 39#>#> [1] 81#>#> An object of class methrix #> n_CpGs: 662 #> n_samples: 4 #> is_h5: FALSE #> Reference: hg19