ibds vs dupcheck

[Mark]

Hi 

I'm trying to identify samples within a dataset, which might be the same individual (the same person my have entered two different studies, for example). 

dupcheck seems to run quicker than idbs, and I was wondering about the relative merits of both. 

Why should I run one over the other? 

Is it possible that dupcheck could find nothing where idbs might?

Thanks

Mark

[Kevin Jacobs <jacobs@bioinformed.com>]

If your goal is simply to detect duplicate samples, then dupcheck is the more efficient approach.  Ibds assumes a population frequency model and is uses it to detect more distant relatives, though in a somewhat crude way.  There are more specialized packages that will do a better job.

Dupcheck and ibds will both find samples that look like MZ-twins, i.e. where concordance is nearly 100% or IBS2 and IBD2 are close to 1.  ibds will often find more than dupcheck, e.g. parent-offspring pairs (IBD=~0, IBD1=IBD2=~0.5), sib-pairs (IBD0=IBD2=~0.25, IBD1=~0.5), etc.

Hope this helps,

-Kevin

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

[Mark]

Great, thanks

Mark

[Mark]

Is there any way to do dupcheck without merging datasets?

For example, I have dataset A which I know contains unique samples, and dataset B, which also contains unique samples, but I want to see if there are any samples in A that are also in B. 

Merging and doing qc.dupcheck would perform a lot of unnecessary comparisons. 

qc.concordance looks like it might do what I want, but I don't quite understand the syntax. 

Thanks

Mark

On Wednesday, 2 April 2014 13:16:42 UTC+1, bioinformed wrote: