PEDCHECK example - Handout from 10/18/2000 course

This handout describes how to use PEDCHECK to check your input pedigree files.

PEDCHECK is distributed as an executable file. The name of the program (as used on the command line) will differ depending on what type of computer you use.

I am preparing this example on a Solaris computer, so my executable is called pedcheck_sol.

Suppose the post-preplink locus file is called datain.dat and the post-makeped pedigree file is called pedin.dat.

Then run:

  pedcheck_sol -d datain.dat -p pedin.dat -4 > errorfile

The flag

specifies that you want the 4th level of error checking.

The filename after the > is where the output gets printed.

If there are no inconsistencies in your data, the bottom of errorfile will look like:

 PedCheck has found NO inconsistencies in the pedigree data.

 If you use PedCheck in any published work, please cite:

 PedCheck: A Program for Identification of Genotype 
 Incompatibilities in Linkage Analysis,
 O'Connell JR, Weeks DE, Am J Hum Genet 63:259-266

If you got some errors, a little further from the bottom you will see something like:


! Summary of Errors: By Pedigree                !
! Pedigree 20                                   !
!        marker M_3                             !
! Pedigree 22                                   !
!        marker M_3                             !
!                                               !
! Summary of Errors: By Marker                  !
! Marker M_3                                    !
!        Pedigree 20                            !
!        Pedigree 22                            !

This indicates that two pedigrees have errors and all the errors involve marker M_3. If you do not put names for your markers, PEDCHECK assigns them names
M_1, M_2, M_3,....
not including the disease locus. Since the LINKAGE format specifies the disease locus as locus 1, the M_1 corresponds to the second locus, but first marker.

If the errors are not too numerous or intertwined for pedigree, marker pair, PEDCHECK will attempt to describe the error. For example


##### GENOTYPE ERROR: Pedigree 20   Locus  3   Name M_3 #####

 ERROR: Child 3 and Father are inconsistent.

 Father 1:  1/1   Mother 2: 3/1   
 Child 3:  3/3   
 Child 4:  3/1   
 Child 5:  1/1   


If the pedigree, marker pair can be made consistent by changing 1 genotype, then PEDCHECK will tell you all the single genotypes that can be changed to restore consistency.
PEDCHECK will tell you what values are consistent.

 Untyping any person listed will result in a consistent pedigree at the given lo
 Using datafile allele frequencies.  

 ##### Pedigree: 20 #######

 Name: M_3   Locus  3. 
                                 (LEVEL 4)
                                 Valid Typings   Odds Against (vs. best) 
 Person 1:   1/  1 (ORIGINAL)
                                    1/  3                    2.000 
 Person 3:   3/  3 (ORIGINAL) 
                                    1/  1                    1.000 
                                    1/  3                    1.000 


As you can see, PEDCHECK suggests two different approaches to resolving the inconsistency:

A. Changing the genotype of one individual to 0 0 (i.e. "untyping")
B. Changing the genotype of one individual to a no-zero value

I very, very strongly recommend doing option A and not option B. Even better would be to redo the genotypes in the laboratory for this pedigree, marker pair.

When it is not possible to restore consistency by changing one individual's genotype, this part of the PEDCHECK output is omitted.

If you have inconsistencies in your data:

  1. Zero out what you believe are the problematic entries in the pre-makeped pedigree file.
  2. Rerun makeped.
  3. Then rerun PEDCHECK

until there are no inconsistencies remaining.