% 
% Input:	
%		haplocells -  haplotype_string array in cells as cgtcangcttac
%		pids - pattern id (not unique) 501, 502, 502, ...
% Output:	
%		ctables - contingency tables snplength cells of #patterns x 2		
%
%		each ctable 
%			a	g	
%		500	4	6
%		501	5	5
%		502	3	2	with missing
%
%		a-97, c-99, g-103, n-110, t-116
% Modified:
% Date: 5/5/2006


function modetable = modegen(haplocells, ctables)

	sm = length(haplocells);	%number of snps in raw data
	sn = length(haplocells{1});	%length of each snp
	snpmat = zeros(sm, sn);
	
	for i=1:sm
		for j=1:sn
			snpmat(i,j) = haplocells{i}(j);
		end
	end

	modetable = snpmat;
	
	for j=1:sn
		ctable = ctables{j};
		rowsum = sum(ctable(2:end, :));
		total = sum(rowsum);

		if rowsum(1) >= rowsum(2)
			modes = ctable(1,1);
		else	
			modes = ctable(1,2);
		end

		if total == 20		%no missing values
			for i=1:sm
				if snpmat(i,j) == modes
					modetable(i,j) = 1;
				else
					modetable(i,j) = 2;
				end
			end
		else
			tot12 = 0;
			for i=1:sm
				if snpmat(i,j) == modes
					modetable(i,j) = 1;
					tot12 = tot12 + 1;
				elseif snpmat(i,j) == 110
					modetable(i,j) = 0;
				else
					modetable(i,j) = 2;
					tot12 = tot12 + 2;
				end
			end
			avg = tot12 / total;
			for i=1:sm
				if snpmat(i,j) == 110;
					modetable(i,j) = avg;
				end
			end
		end

	end