Generate a matrix with the probabilities of observed genotypes (columns) conditional on actual genotypes (rows), or return a function to generate such matrices (using a single value Err as input to that function).
ErrToM(Err = NA, flavour = "version2.0", Return = "matrix")estimated genotyping error rate, as a single number or 3x3 or 4x4
matrix. If a single number, an error model is used that aims to deal with
scoring errors typical for SNP arrays. If a matrix, this should be the
probability of observed genotype (columns) conditional on actual genotype
(rows). Each row must therefore sum to 1. If Return='function', this
may be NA.
matrix-generating function, or one of 'version2.0',
'version1.3' (='SNPchip'), 'version1.1' (='version111'), referring to the
sequoia version in which it was used as default. Ignored if Err is a
matrix and Return='matrix' (in which case the matrix will only be
checked for validity).
output, 'matrix' (always 3x3) or 'function'.
Either a 3x3 matrix, or a function generating a 3x3 matrix.
By default (flavour = "SNPchip"), Err is interpreted
as a locus-level error rate (rather than allele-level), and equals the
probability that an actual heterozygote is observed as either homozygote
(i.e., the probability that it is observed as AA = probability that
observed as aa = Err/2). The probability that one homozygote is
observed as the other is (Err/2\()^2\).
The inbuilt 'flavours' correspond to the presumed and simulated error structures, which have changed with sequoia versions. The most appropriate error structure will depend on the genotyping platform; 'version0.9' and 'version1.1' were inspired by SNP array genotyping while 'version1.3' and 'version2.0' are intended to be more general.
Pr(observed genotype (columns) | actual genotype (rows)):
version2.0:
| 0 | 1 | 2 | |
| 0 | \((1-E/2)^2\) | \(E(1-E/2)\) | \((E/2)^2\) |
| 1 | \(E/2\) | \(1-E\) | \(E/2\) |
| 2 | \((E/2)^2\) | \(E(1-E/2)\) | \((1-E/2)^2\) |
version1.3
| 0 | 1 | 2 | |
| 0 | \(1-E-(E/2)^2\) | \(E\) | \((E/2)^2\) |
| 1 | \(E/2\) | \(1-E\) | \(E/2\) |
| 2 | \((E/2)^2\) | \(E\) | \(1-E-(E/2)^2\) |
version1.1
| 0 | 1 | 2 | |
| 0 | \(1-E\) | \(E/2\) | \(E/2\) |
| 1 | \(E/2\) | \(1-E\) | \(E/2\) |
| 2 | \(E/2\) | \(E/2\) | \(1-E\) |
version0.9 (not recommended)
| 0 | 1 | 2 | |
| 0 | \(1-E\) | \(E\) | \(0\) |
| 1 | \(E/2\) | \(1-E\) | \(E/2\) |
| 2 | \(0\) | \(E\) | \(1-E\) |