Counting the unique values

I have a values

out=

          c1       c2,,,,,,,,,,,,,,,,,,,,,c5 

'gene1' 'd'  'u'  'd'  'u'  'd'  
'gene2' 'u'  'u'  'u'  'u'  'd'
'gene3' 'u'  'u'  'd'  'u'  'u'
'gene4' 'd'  'u'  'u'  'd'  'd'
'gene5' 'u'  'u'  'u'  'u'  'd'
'gene6' 'd'  'u'  'u'  'u'  'u'
'gene7' 'd'  'u'  'd'  'u'  'u'

Taking the first column 'c1' value for gene1 is 'd' this value must be compared with all other column if there is more than 3 same vales it should be displayed,,Only first clumn must be compared with others columns

in 1st col there are 3d's ,2nd col 4u's(since 1st col is u),3rd col

4u's,,,,,,,,,,,6&7th gene's there are only 1 ans 2 d's respectively,so it should be deleted

So i need output as

'gene1' 'd'  'u'  'd'  'u'  'd'  
'gene2' 'u'  'u'  'u'  'u'  'd'
'gene3' 'u'  'u'  'd'  'u'  'u'
'gene4' 'd'  'u'  'u'  'd'  'd'
'gene5' 'u'  'u'  'u'  'u'  'd'

Pleae help

[Jan]

@Pat: Please post the input data in valiud Matlab syntax, such that we can try our suggestion by copy&paste. The time required to guess what "c1 c2,,,,,c5" should mean is wasted, because you know this detail.

Answer 1

[Freddy]

Hello Pat,

first idea i came up with:

A = {'','c1','c2','c3','c4','c5'; 
  'gene1' 'd'  'u'  'd'  'u'  'd';...
  'gene2' 'u'  'u'  'u'  'u'  'd';...
  'gene3' 'u'  'u'  'd'  'u'  'u';...
  'gene4' 'd'  'u'  'u'  'd'  'd';...
  'gene5' 'u'  'u'  'u'  'u'  'd';...
  'gene6' 'd'  'u'  'u'  'u'  'u';...
  'gene7' 'd'  'u'  'd'  'u'  'u'};
limit = 3;
F = cell2mat(A(2:end,2:end));
A(logical([0;sum(bsxfun(@eq,F,F(:,1)),2)>=limit]),:);

Hopefully it will help you.

Freddy

[Jan]

You can omit the "logical", if you use:

A([true; sum(bsxfun(@eq, F, F(:,1)), 2) >= limit], :)

Answer 2

[Walter Roberson]

c1_column = 2;  %looks like column 2 to me, since column 1 has gene name

match_count = arrayfun(@(K) sum(out{K,cl_column} == [out{K,c1_column+1:end}]), 1:size(out,1));
out(match_count > 3, :)

Your problem description is inconsistent about what to do if the number of matches is exactly 3. You wrote that it has to be more than 3, but your sample output includes the case where it is exactly 3.

[Pat]

Walter thats the sample program,in your code i get error as

Undefined function or variable 'cl_column'.

[Jan]

@Pat: Come on, I'm sure you are able to fix this typo by your own. The "1" (one) looks very similar to the "l" (lowercase L). It is your job to participate as far as possible in the solution of your problems.

Answer 3

[Bjorn Gustavsson]

Pat,

I strongly suggest you change the encoding of your data! Make your variable an integer array with for example 0 for 'd' and 1 for 'u'. Then you could do something like this:

genes = [0,1,0,1,0
         1 1 1 1 0
         1 1 0 1 1
         0 1 1 0 0]; 
lim4disp = 3;
genes(sum(repmat(genes(:,1),[1,size(genes,2)]) == genes,2)>=lim4disp,:)