Replaces missing marker scores with the mode scores of the most similar genotypes (L.C.P. Keizer, J.T.N.M. Thissen & F.A. van Eeuwijk).
Options
PRINT = string tokens |
What to print (summary, similarity, neighbours, details); default summ |
|---|---|
NNEIGHBOURS = scalar |
Number of nearest neighbours; default 5 |
MAXDISTANCE = scalar |
Maximum similarity difference; default 0.1 |
Parameters
MKSCORES = pointers |
Pointer with the original marker scores; must be set |
|---|---|
MKNAMES = texts |
Marker names |
IDMGENOTYPES = texts |
Labels for genotypes |
NEWMKSCORES = pointers |
Pointer to store the new marker scores; must be set |
Description
QMVREPLACE replaces missing marker scores with the mode score of the most similar genotype(s). The marker scores with missing values are supplied by the MKSCORES pointer, which contains a factor for each marker. The length of factors is the number of genotypes. The new factors, in which the missing marker scores are replaced, can be saved by the NEWMKSCORES pointer. The MKNAMES and IDMGENOTYPES parameters can be set to obtain more readable output.
QMVREPLACE forms a similarity matrix from the marker scores using the FSIMILARITY directive with parameter TEST=simplematching. The NNEIGHBOURS option specifies the number of most-similar neighbouring genotypes to use when filling in the missing values for each genotype (default 5). To prevent the use of neighbours that are too different from the genotype, neighbours are selected only if their distances from the genotype are less than the value supplied by the MAXDISTANCE option (default 0.1). For each missing marker score of the genotype, a replacement value is obtained by taking the score that is most common amongst the selected neighbouring genotypes (i.e. the mode of their values). If the marker scores of the closest genotypes are all missing, the missing value is not replaced.
The PRINT option controls the printed output with settings:
summary |
prints general information about the replaced marker missing scores, |
|---|---|
similarity |
prints the similarity matrix, |
neigbours |
prints the most-similar neighbours of the genotypes with missing marker scores, and |
details |
prints information about each replacement, |
Options: PRINT, NNEIGHBOURS, MAXDISTANCE.
Parameters: MKSCORES, MKNAMES, IDMGENOTYPES, NEWMKSCORES.
Action with RESTRICT
Restrictions are not allowed.
See also
Procedures: MULTMISSING, QMVESTIMATE, SVHOTDECK.
Commands for: Statistical genetics and QTL estimation.
Example
CAPTION 'QMVREPLACE example'; STYLE=meta
QIMPORT [POPULATION=amp] '%GENDIR%/Examples/Qassociation_geno.txt';\
MAPFILE='%GENDIR%/Examples/Qassociation_map.txt';\
MKSCORES=scores; CHROMOSOMES=mkchr; POSITIONS=mkpos; MKNAMES=mknames
QMVREPLACE [PRINT=summary] scores; NEWMKSCORES=mkout