familycheck: Error check for a family classification

Description

Given a family id vector, also compute the familial grouping from first principles using the parenting data, and compare the results.

Usage

familycheck(famid, id, father.id, mother.id, newfam)

Arguments

famid

a vector of family identifiers

a vector of unique subject identifiers

father.id

vector containing the id of the biological father

mother.id

vector containing the id of the biological mother

newfam

the result of a call to makefamid. If this has allready been computed by the user, adding it as an argument shortens the running time somewhat.

Value

a data frame with one row for each unique family id in the famid argument. Components of the output are
famidthe family id, as entered into the data set
nnumber of subjects in the family
unrelatednumber of them that appear to be unrelated to anyone else in the entire pedigree set. This is usually marry-ins with no children (in the pedigree), and if so are not a problem.
splitnumber of unique "new" family ids. If this is 0, it means that no one in this "family" is related to anyone else (not good); 1 = everythings is fine; 2+= the family appears to be a set of disjoint trees. Are you missing some of the people?
joinnumber of other families that had a unique famid, but are actually joined to this one. 0 is the hope. If there are any joins, then an attribute "join" is attached. It will be a matrix with famid as row labels, new-family-id as the columns, and the number of subjects as entries.

Details

The makefamid function is used to create a de novo family id from the parentage data, and this is compared to the family id given in the data.

Examples

Run this code

#
# This is from a pedigree that had some identifier errors
#
> checkit<- familycheck(ids2$famid, ids2$gid, ids2$fatherid, ids2$motherid)
> table(checkit$split)    # should be all 1's
   0   1 2 
 112 424 4
# Shows 112 of the "families" were actually isolated individuals,
#  and that 4 of the families actually split into 2.  
#  In one case, a mistyped father id caused one child, along with his spouse
#  and children, to be "set adrift" from the connected pedigree.

> table(checkit$join)
   0 1 2 
 531 6 3
#
# There are 6 families with 1 other joined to them (3 pairs), and 3 with
#  2 others added to them (one triplet).
# For instance, a single mistyped father id of someone in family 319,
#  which was by bad luck the id of someone else in family 339,
#  was sufficient to join two groups.
> attr(checkit, 'join')
    [,1] [,2] [,3] [,4] [,5] [,6] [,7] 
 31   78    0    0    0    0    0    0
 32    3   15    0    0    0    0    0
 33    6    0   12    0    0    0    0
 63    0    0    0   63    0    0    0
 65    0    0    0   17   16    0    0
122    0    0    0    0    0   16    0
127    0    0    0    0    0   30    0
319    0    0    0    0    0    0   20
339    0    0    0    0    0    0   37