tag_mendelian_incompatibilities.RdThis is a handy function for taking parent-offspring pairs, typically identified using CKMRsim, and then tagging the loci at which the observed/recorded genotypes are not compatible with Mendelian inheritance. This version is currently very simple and merely looks at pairs of individuals. (i.e., it does not attempt to identify Mendelian incompatible loci on the basis of pa-ma-kid trios.) Noetheless, it should be useful for identifying poor-performing loci.
tag_mendelian_incompatibilities(po, geno)a tibble of the parent-offspring pairs. This will typically be obtained
by filtering the output for pairwise_kin_logl_ratios() and it must have the format
of that output: columns of D2_indiv, D1_indiv, logl_ratio, and num_loc.
a long format data frame / tibble of genotype data for both the candidate
offspring and the candidate parents. Typically this is what you get if you use
bind_rows() on the D1 and D2 inputs to pairwise_kin_logl_ratios() for a parent-offspring
type of analysis. Must have the
columns: Indiv, Locus, gene_copy, and Allele. Obviously, the IDs in Indiv must
correspond to those used in D1_indiv and D2_indiv
This returns a tibble with the columns:
D2_indiv: the ID of the D2_indiv
D1_indiv: the ID of the D1_indiv
logl_ratio: the log-likelihood ratio for parent-offspring for the pair
Locus: the name of the locus
D1a_1, D1a_2, D2a_1, D2a_2: D1_indiv's allelic type of gene copy 1 and 2, and then
D2_indiv's allelic type of gene copy 1 and 2, respectively
is_MI: logical vector. TRUE if the pair is Mendelian incompatible at the Locus. Otherwise
FALSE (or NA if either member of pair was NA)
Note that if you want to count up the relative frequency (across all pairs) of Mendelian incompatibilities at each locus you can do like this:
tag_mendelian_incompatibilities(po, geno) %>%
group_by(Locus) %>%
summarise(mean_incompat = mean(is_MI, na.rm = TRUE)) %>%
arrange(desc(mean_incompat))