refine.qtl()

[Joshua Havill]

Hello Karl et al.,

I'm working within a four-way cross (4WC) framework, and have identified QTL for various traits. Since CIM and MQM are not built to work within the 4WC framework, I have been using scanone() to identify putative QTL positions.

As a follow-up, I have been trying to refine the positions of said QTL by using the refine.qtl() position. In some cases, the position changes but only slightly. In other cases with additional traits, it appears to shift to a different position entirely, so I'm not sure about how to proceed with the interpretation or what potential issues are causing the "instability" of the identified QTL.

I see the possibility of multiple QTL being present, but my population isn't large enough to detect small or even mid-size effect QTL that may be segregating (realistically) in the presence of other major (and potentially linked) QTL.

I'm wondering if anyone has any input into what may be causing said patterns of shift following the refine.qtl() process.

Thanks in advance,

Josh

[Karl Broman]

Could you give more information about what you mean by the results being unstable?

I don't quite know how to say.

refineqtl takes a given multiple QTL model and tries to optimize the locations of the QTL. The number of QTL on each chromosome is kept fixed, and it iteratively scans each QTL location, keeping all others in fixed positions, and moves the location if it finds any improvement in log likelihood.

With a single-QTL model, the results should be identical to scanone(). With a two-QTL model, the results should be identical to scantwo().

(Assuming the method="hk"/"imp" and model="normal"/"binary" are the same.)

karl

[Joshua Havill]

Hi Karl,

Let's say I run scanone(), makeqtl(), and fitqtl(). If I run refine.qtl() using the qtl.object and the single qtl formula, the position shifts, sometimes a couple of centimorgans, other times it shifts to the opposite end of the linkage group, depending on the trait.

Yes, both "hk" and "binary" are consistently called as needed between scanone() and fitqtl().

--
You received this message because you are subscribed to a topic in the Google Groups "R/qtl discussion" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/rqtl-disc/fBTK9RFxqCM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rqtl-disc+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rqtl-disc/1e11bd68-0e32-4458-bcfd-edfd0bca44afn%40googlegroups.com.

--

Joshua Havill, M. Sc.
C: (612)-836-7834

O: (612)-625-5274

Graduate Research Assistant, Ph. D. Candidate

Plant and Microbial Biology Graduate Program

Muehlbauer Lab

Department of Agronomy and Plant Genetics

411 Borlaug Hall

1991 Upper Buford Circle

Saint Paul, MN 55108

University of Minnesota - Twin Cities

[Joshua Havill]

Hi Karl,

Nevermind, I made some typos. Yes, the position stays the same. Thanks for your e-mail.

Josh