Date:         Mon, 31 Oct 2011 18:25:30 -0700
Reply-To:     oslo <oslo@yahoo.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         oslo <hokut1@YAHOO.COM>
Subject:      Re: commen frequency
Comments: To: Tom Abernathy <tom.abernathy@GMAIL.COM>
In-Reply-To:  <201110312324.p9VIZHZn027626@waikiki.cc.uga.edu>
Content-Type: text/plain; charset=iso-8859-1

Dear Tom;
 
Your helps are sincerely appreciated,
 
 
Thanks.
 
Oslo


________________________________
From: Tom Abernathy <tom.abernathy@GMAIL.COM>
To: SAS-L@LISTSERV.UGA.EDU
Sent: Monday, October 31, 2011 6:24 PM
Subject: Re: commen frequency

On Mon, 31 Oct 2011 14:27:56 -0700, oslo <hokut1@YAHOO.COM> wrote:

>Tom;

Thanks a lot. I think I could not explain the problem correctly. The
variables (rsxxxxxxxx) are unique in each column. That is there is only one
rs29024430 in column A. Likewise there is only one rs29009912 in the same
column (A). This is the case for other columns too. So overall I need the
get the similarity between columns. That is how many rsxxxxxxxx are common
in both column A and B. In other words how many of rsxxxxxxxx which exist in
column A are also in column B and so on.

Regards,

Oslo

So merge them by gene and set a 0/1 variable for each of the four sources.
Then create the combinations using AND.  Something like this.

data current;
length a b c d $10.;
input a--d;
cards;
rs29024430  rs43708440  rs29024430   rs29024430
rs29009912  rs29009907  rs43708440   rs43708440
rs29009979  rs29009912  rs29009907   rs29010802
rs29010147  rs29012086  rs29009912   rs17871338
rs29010295  rs29012174  rs29009979   rs29009907
rs29010510  rs29013844  rs29010295   rs29009912
rs29011155  rs29016146  rs29010461   rs29009979
rs29012070  rs29017980  rs29011155   rs29010295
rs29012174  rs29018185  rs29012070   rs29010461
;;;;

proc sql noprint ;
  create table a as select 1 as a,a as gene from current order by gene;
  create table b as select 1 as b,b as gene from current order by gene;
  create table c as select 1 as c,c as gene from current order by gene;
  create table d as select 1 as d,d as gene from current order by gene;
quit;

data want;
  merge a b c d;
  by gene;
  a = sum(0,a);
  b = sum(0,b);
  c = sum(0,c);
  d = sum(0,d);
  ab = a and b;
  ac = a and c;
  abc = a and b and c;
run;

proc means;
run;