Detect deviants from SNPs; classify SNPs

Detect deviant SNPs using excess of heterozygotes (alleles that do not follow HWE) and allelic-ratio deviations (alleles with ratios that do not follow a normal Z-score or chi-square distribution). See details.

Usage

dupGet(
  data,
  test = c("z.het", "z.05", "z.all", "chi.het", "chi.05", "chi.all"),
  intersection = FALSE,
  method = c("fisher", "chi.sq"),
  plot = TRUE,
  verbose = TRUE,
  ...
)

Arguments

data: data frame of the output of allele.info
test: character. type of test to be used for significance. See details
intersection: logical, whether to use the intersection of the methods specified in test (if more than one)
method: character. method for testing excess of heterozygotes. Fisher exact test (fisher) or Chi-square test (chi.sq)
plot: logical. whether to plot the detected singlets and duplicates on allele ratio vs. proportion of heterozygotes plot.
verbose: logical. show progress
...: additional parameters passed on to plot

Value

Returns a data frame of snps/alleles with their duplication status

Details

SNP deviants are detected with both excess of heterozygosity according to HWE and deviant SNPs where depth values fall outside of the normal distribution are detected using the following methods:

Z-score test \(Z_{x} = \sum_{i=1}^{n} Z_{i}\); \(Z_{i} = \frac{\left ( (N_{i}\times p)- N_{Ai} \right )}{\sqrt{N_{i}\times p(1-p)}}\)
chi-square test \(X_{x}^{2} = \sum_{i-1}^{n} X_{i}^{2}\); \(X_{i}^{2} = (\frac{(N_{i}\times p - N_{Ai})^2}{N_{i}\times p} + \frac{(N_{i}\times (1 - p)- (N_{i} - N_{Ai}))^2}{N_{i}\times (1-p)})\)

See references for more details on the methods

Users can pick among Z-score for heterozygotes (z.het, chi.het), all allele combinations (z.all, chi.all) and the assumption of no probe bias p=0.5 (z.05, chi.05)

Author

Piyal Karunarathne

Examples

if (FALSE) data(alleleINF)
DD<-dupGet(alleleINF)
#> categorizing deviant SNPs with 
#>  excess of heterozygotes z.all & chi.all