Building Linkage Map From Lod Scores Between Snp Markers

Question

[Eric Normandeau]

Hi,

I have a dataset consisting of pairs of SNP loci with LOD (logarithm of the odds) scores representing the chance that these be physically linked in the genome based of recombination frequencies. Here is what the file looks like (spaces represent tabulations):

locus_01    locus_02    45.32
locus_01    locus_65    38.86
locus_05    locus_23    35.60
locus_23    locus_41    32.09
locus_65    locus_47    27.13
...
locus_34    locus_65    3.01

The file is close to 1000 lines and gives information for about 250 markers and their association (LOD score) between them, with a minimal LOD score of 3. The markers included in this file have been tested for Hardy Weinberg disequilibrium and were ok in this regard.

What other a priori test would you consider?

I would like to build the linkage groups containing these markers in an automated way since the number of markers may be well above 1000 in incoming datasets.

What software would you recommend to build such a map?

I implemented a program to do just that, but ended up thinking that there were many things one should be careful about when building a linkage map and that many of these were not tested by my small program. What would you use? Ideally, the program should be usable from the terminal under Linux.

Cheers!

Answer 1

[Jashapiro]

You might look at the tools available in R/qtl. There are a variety of tools starting from the genotypes of individuals to do all of the HW checks and LOD scores (which you have already done), through the generation of the actual map. The whole process is outlined in this tutorial (pdf link), including investigation of problematic markers to detect errors in the map.

[Eric Normandeau]

Hi @JAShapiro, I'll be sure to have a look at this package! Thanks