% splits deCODEme_csv.csv file into a separate csv file for each chromosome
% 
% the file takes some time to run, as it is a big file with strings
% note: pay attention on whether the file has a header line or not

%read various columns of csv file as string
disp('Reading deCODEme_scan.csv')
fi=fopen('deCODEme_scan.csv');
%if the csv file does not contain a header line
decodeme=textscan(fi, '%s %s %s %s %s %s', 'delimiter',',');
%if it does, instead use this command
%decodeme=textscan(fi,'%s %s %s %s %s %s', 'delimiter',',','headerLines',1);
fclose(fi);
%decodeme{i} is a vector of strings
ndecodeme=size(decodeme{1},1);
disp('deCODEme.csv read');

%find where  chr X,Y and M begins
i=1;
while ~(strcmp(decodeme{3}(i),'X'))
      i=i+1;
end
initX=i;
while ~(strcmp(decodeme{3}(i),'Y'))
      i=i+1;
end
initY=i;
while ~(strcmp(decodeme{3}(i),'M'))
      i=i+1;
end
initM=i;
%note. format data
%rs4477212,A/G,1,72017,+,AA

%extract files for chr 1:22
disp('Start extracting')
chromosome=str2num(char(decodeme{3}(1:initX-1,:)));
for i=1:22
    disp(strcat('Extracting chr',num2str(i)))
    fid=find(chromosome==i);
    n_rs_chr=size(fid,1);
    filetosavestring=strcat('cacio',num2str(i),'.csv');
    fi=fopen(filetosavestring,'w');
    for j=1:n_rs_chr
    	mystring=decodeme{1}(fid(j));
	for k=2:6
    	    mystring=strcat(mystring,',',decodeme{k}(fid(j)));
	end
    towrite=char(mystring);
    fprintf(fi,'%s\n',towrite);    	
    end
    fclose(fi);
    disp(strcat('Done - extracted chr',num2str(i)))
end

%extract X,Y,M
disp('Extracting chr X')
fi=fopen('cacioX.csv','w');
for i=initX:initY-1
    mystring=decodeme{1}(i);
    for j=2:6
    	    mystring=strcat(mystring,',',decodeme{j}(i));
    end
    towrite=char(mystring);
    fprintf(fi,'%s\n',towrite);   
end
fclose(fi);
disp('Done - extracted chr X')
disp('Extracting chr Y')
fi=fopen('cacioY.csv','w');
for i=initY:initM-1
    mystring=decodeme{1}(i);
    for j=2:6
    	    mystring=strcat(mystring,',',decodeme{j}(i));
    end
    towrite=char(mystring);
    fprintf(fi,'%s\n',towrite);   
end
fclose(fi);
disp('Done - extracted chr Y')
disp('Extracting mtDNA')
fi=fopen('cacioM.csv','w');
for i=initM:ndecodeme
    mystring=decodeme{1}(i);
    for j=2:6
    	    mystring=strcat(mystring,',',decodeme{j}(i));
    end
    towrite=char(mystring);
    fprintf(fi,'%s\n',towrite);   
end
fclose(fi);
disp('Done - extracted mtDNA')