Question on plots of recomb freq and lod

[HASmith]

Hi,
I'm trying to use rqtl to assess recombination frequency. Yet when I use the plotRF function, I know that red = high lod score & low recombination frequency, but what is the scale (i.e., what cutoffs/numeric thresholds are used to define red, blue, etc.)?

For me an even better result would be perhaps if it were possible to generate a plot showing all markers on the x-axis, and the recombination frequency on the y-axis. I found a somewhat similar function described in an online rqtl guide to mapping (rf <- pull.rf(mapthis); lod <- pull.rf(mapthis, what="lod"); plot(as.numeric(rf), as.numeric(lod), xlab="Recombination fraction", ylab="LOD score"). Yet as this is plotting against lod score, it's not quite what I need. The closest I came up with was to use write.csv to save the "rf" matrix to a file, then take the average down each column for a value of the recombination frequency at that marker. Does this sound correct, or is there a better way? Sorry if it's a somewhat basic question.

Thank you in advance,
Hilary

[Karl Broman]

On Jan 28, 2013, at 10:25 AM, HASmith <hasmith...@gmail.com> wrote:

> Hi,
> I'm trying to use rqtl to assess recombination frequency. Yet when I use the plotRF function, I know that red = high lod score & low recombination frequency, but what is the scale (i.e., what cutoffs/numeric thresholds are used to define red, blue, etc.)?

The scale is relatively arbitrary. We threshold the high LOD scores (argument zmax, with default value 10) and transform the recombination fractions to be on a similar scale (with the formula -4*(log2(rf) + 1)/12 * zmax ).

> For me an even better result would be perhaps if it were possible to generate a plot showing all markers on the x-axis, and the recombination frequency on the y-axis. I found a somewhat similar function described in an online rqtl guide to mapping (rf <- pull.rf(mapthis); lod <- pull.rf(mapthis, what="lod"); plot(as.numeric(rf), as.numeric(lod), xlab="Recombination fraction", ylab="LOD score"). Yet as this is plotting against lod score, it's not quite what I need. The closest I came up with was to use write.csv to save the "rf" matrix to a file, then take the average down each column for a value of the recombination frequency at that marker. Does this sound correct, or is there a better way? Sorry if it's a somewhat basic question.

I'm not sure what you're trying to look at. There is an estimated recombination fraction and LOD score for every pair of markers. The code you cite, using pull.rf(), pulls out the two things as two separate square matrices, with rows and columns corresponding to markers. But I'm not sure why you'd be interested in the column averages.

karl

[HASmith]

Thanks; I was simply looking for a way to calculate summary statistics about recombination rates. I think I found how to obtain the information we needed using the est.map to convert our prior physical-distance map to one in cM, and then taking that estimate of genetic distance (chromosome length in cM) divided by physical distance (Megabases) to yield a recombination rate for the chromosome. We're working with insects and I just used the Haldane map function; if I recall for this purpose map function wouldn't matter much.

I was wondering about taking column averages from the pull.rf() as a way to convert from a pairwise recombination rate, to an average recombination rate for each marker. So far I've just ended up taking cM/Mb, or, I think equivalently (values are very close albeit not exactly the same), computing:
(100 * output of count.xo) / (total # of individuals) = recombination rate in cM/Mb.
Thanks again,
Hilary

[Karl Broman]

If your interest is in estimating recombination rates, I don't see how the pairwise recombination fractions from est.rf will be of any use to you.

My package R/xoi includes a function est.recrate() that uses a genetic map and physical map of a chromosome to get a smoothed estimated of recombination rate along the chromosome.

https://github.com/kbroman/xoi
http://cran.r-project.org/web/packages/xoi/

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?hl=en.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

[HASmith]

Thank you very much; this looks really useful. If there's any associated publication that I could cite for this program/it's use, please let me know.