dominant markers and formLinkageGroups()

[Jen]

Hello,

I am trying to build a linkage map from an F2 population genotyped with a mix of co-dominant and dominant markers. I have coded the data appropriately as is described in read.cross(), however I am concerned that the dominant markers are still not being handled properly. 

Specifically, I am running into problems when I attempt to form the linkage groups. When I try to form linkage groups using formLinkageGroups with the full dataset, no matter how I set the parameters max.rf and min.lod, essentially all of the markers come out as linked on the first chromosome -- I know this is not correct, as I expect to see 22 chromosomes for this species. 

Interestingly, when I separate the co-dominant markers from the dominant markers, the co-dominant markers sort themselves out into approximately the right number of linkage groups, while the dominant markers all come out as linked on a single chromosome. Furthermore, I get a warning message on loading the file containing only dominant markers: "strange genotype pattern". The genotype pattern is not strange for dominant markers, it is exactly what you would expect (I know because I ran a chi-square test). 

Do you know why I might be seeing a., this warning message, and b. this problem forming linkage groups with the dominant markers?

Thanks very much,

Jen

[Karl Broman]

It's hard to tell without looking at the data.
I've not seen this problem before.

My guess is that the genotypes at dominant markers weren't interpreted correctly.

Use geno.table() and check that the genotype codes look to be correct.

karl

> --
> You received this message because you are subscribed to the Google Groups "R/qtl discussion" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to rqtl-disc+...@googlegroups.com.
> To post to this group, send email to rqtl...@googlegroups.com.
> Visit this group at http://groups.google.com/group/rqtl-disc.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[Jen]

Thanks very much for getting back to me. I ran geno.table() and the results were as I expected, with the exception of of the p-value column for the chi-square -- I'm assuming that is because the chi square is only comparing the values for AA, AB, and BB, and not not.BB and not.AA

The first three dominant markers are illustrative of all the dominant markers in the output of geno.table():

                     chr missing  AA   AB  BB   not.BB   not.AA      P.value

DR_09180       1      22        0     0    0        197         61         0.000000e+00

DR_03566       1       6         0     0    0        56          218        0.000000e+00

DR_08039       1      20        0     0    0        58          202        0.000000e+00

The codominant markers look like the following three examples, although I do have some cases where not.BB and not.AA are nonzero for some of the codominant markers due to missing allele data. 

                   chr   missing   AA  AB    BB   not.BB   not.AA      P.value

*AA003N        1       4         71   139    66      0           0            9.068066e-01

*AA015N        1       1         77   128    74      0           0            3.752035e-01

*AA035N        1       3         71   150    56      0           0            1.708195e-01

Can you think of any other reasons why I might be experiencing a problem mapping the dominant markers? 

Thanks very much,

Jen

[Karl Broman]

The chi-square tests in geno.table, in the presence of dominant genotypes, assume that the "not BB" genotypes would be accompanied by "BB" genotypes, in the proportion 3:1. Similarly, we expect that the "not AA" genotypes would be accompanied by "AA" genotypes, in the proportion 3:1.

Seeing only "not AA" and "not BB" but no "AA" or "BB" seems really unusual.

karl

[Jen]

Ah hah -- I see now how we made a mistake with the coding. Thanks very much!

[Jen]

Form linkage groups is now working very well since I have fixed my coding, however I now get warnings that markers are not converging when I run est.rf(). This didn't seem to affect the formation of the linkage groups, but I started running into problems when attempting to compare marker orders using ripple().

I was wondering if you could tell me what exactly the non-convergence warning means, and what might be the cause of it?

Thanks very much,

Jen

would a simple solution to this problem be to remove only markers 3 and 4? 

On Thursday, July 11, 2013 12:23:10 AM UTC-4, Karl Broman wrote:

[Karl Broman]

If you use ripple() with method="likelihood", it uses an iterative algorithm to estimate inter-marker distances and evaluate the likelihood for a marker order.

There's a tolerance value to check for convergence (that the estimates haven't changed from one iteration to the next) plus some maximum number of iterations. If the algorithm reaches the maximum number of iterations with the estimates still changing by more than the tolerance value at each iteration, that warning appears.

You can probably ignore it, because the likelihood for each order can be reasonably well established even without reaching convergence.

karl

[Saunak Sen]

One possible reason for the convergence messages is that with dominant markers there is reduced information on recombination relative to co-dominant markers.  For example, if you have a AA/notAA dominant marker next to a co-dominant marker, individuals who are BB for the co-dominant marker are essentially non-informative when the recombination fraction is small.  If you have a AA/notAA marker next to a BB/notBB marker, you have almost no information about recombination between them unless the recombination fraction is large.

Saunak

[mukesh choudhary]

Can you please help me with how the coding was fixed for dominant markers? I am struggling with my InDEL markers (dominant markers) data for F2 population linkage map construction.

Thanks