Date:         Tue, 29 Jul 2008 13:37:45 -0500
Reply-To:     Mary <mlhoward@avalon.net>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Mary <mlhoward@AVALON.NET>
Subject:      Re: data mining
Comments: To: msz03@albany.edu
Content-Type: text/plain; charset="iso-8859-1"

True, but I don't trust that the user asking the question really knows the subject area, particularly when this user is asking for substrings of 5 long when SNP alleles are 2 each, and therefore haplotypes have an even number of alleles!  

That's always the SAS-L dilemma- do you give the user what they asked for, or what you think they should have asked for??

-Mary
  ----- Original Message ----- 
  From: Mike Zdeb 
  To: SAS-L@LISTSERV.UGA.EDU 
  Sent: Tuesday, July 29, 2008 1:19 PM
  Subject: Re: data mining


  hi ... I figured that'd be easy to suppress results that are not 'reasonable' given knowledge of
  the subject area

  I think that the hard part is done here, i.e. finding the strings and repeats in one pass through
  the data, and like I said, it does agree with the posting except for the extra line

  --
  Mike Zdeb
  U@Albany School of Public Health
  One University Place
  Rensselaer, New York 12144-3456
  P/518-402-6479 F/630-604-1475

  > Note that you've got a start position of 16; my solution is assuming that this data is haplotypes,
  > and there's two alleles to every marker, so I'm not including strings that go across markers, this
  > particular SNP marker would be "tc" in columns 15-16.
  >
  > So including that extra results depends on whether you want to do what's appropriate for the field
  > or not- I don't think it is appropriate to report a haplotype string that splits a SNP in
  > genetics, and thus I don't use the strings that start in even columns.   The user doesn't say that
  > this is genetics data, but given the letters used, it is likely.
  >
  > -Mary