xExcludeOverlap(Candidates) removes candidates which overlap with one another, leaving only the one that occurs first in the list
0001 % xExcludeOverlap(Candidates) removes candidates which overlap with one 0002 % another, leaving only the one that occurs first in the list 0003 0004 function [Candidates,Discrepancy] = xExcludeOverlap2(Candidates, Discrepancy, Perm) 0005 0006 N = length(Candidates(1,:)) - 1; % number of nucleotides 0007 0008 Keep = ones(size(Candidates(:,1))); % rows of Candidates which are kept 0009 0010 [C,ord] = sortrows(Candidates,[N+1 Perm]); % sort by filenumber, nucleotides 0011 0012 %D = diff(C); % matrix of row differences 0013 %D(:,1:(N+1)) = abs(sign(D(:,1:(N+1))));% 1 for every difference, 0 if same 0014 0015 a = 1; % index of current lowest discrepancy 0016 for i = 2:(length(C(:,1))-1), % go through row differences 0017 if C(a,N+1) == C(i,N+1), % if file numbers agree 0018 if sum(C(a,1:N) == C(i,1:N)) > N/2, % if more than half agree 0019 if Discrepancy(ord(i)) > Discrepancy(ord(a)), 0020 Keep(i) = 0; % reject i; a has lower discrep 0021 else 0022 Keep(a) = 0; % reject a; i has lower discrep 0023 a = i; % found one with lower discrepancy 0024 end 0025 else 0026 a = i; % moved on to new motif 0027 end 0028 else 0029 a = i; % moved on to new file 0030 end 0031 end 0032 0033 Keep(ord) = Keep; % undo the re-ordering 0034 OK = find(Keep); 0035 Candidates = Candidates(OK,:); % return only non-overlapping candidates 0036 Discrepancy = Discrepancy(OK);