How to copy specific columns of a character array to another character array?

I have a character array named 'seq' as following:

MRSLMVLALLAVAALCLCLAGPADAKPSSAESRKGGATFVSKREGSEVVRRLRRYLDSGL

MRTPMLLALLALAT--LCLAGRADAKPGDAESGK-GAAFVSKQEGSEVVKRLRRYLDHWL

MRALTLLALLALAT--LCITGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL

MRALTLLALLALAA--LCIAGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL

MRALTLLALLALAA--LCIAGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL

I want to remove those columns which has more than 50% gaps (-) in them. For example, column 15 has 4 gaps out of 5, so I want to remove that column. After removing these columns, I want to take the rest to another character array (e.g 'mod_seq').

I tried with this code:

 [rowLen, colLen] = size(seq);
count = 0;
mod_col_no = 1;
mod_seq = zeros(rowLen,colLen);
  end
for i=1:colLen
    for j=1:rowLen
        if seq(j,i) == '-' 
            count = count+1;
        end     
    end
    if count > ceil(rowLen/2) 
         continue;
    else
            mod_seq(:,mod_col_no) = seq(:,i);
            mod_col_no=+1;
    end
end
mod_seq = char(mod_seq);

I have tried like this because I cannot initialize 'mod_seq'. I can't simply make the two arrays equal because when I delete columns, mod_seq will have less dimension.

Answer 1

[Guillaume]

The way to fix your loop would be to gather the list of columns to delete inside the loop and perform the deletion after the loop.

Even simpler is to not bother with loops at all when a single line of code can do the same:

seq=['MRSLMVLALLAVAALCLCLAGPADAKPSSAESRKGGATFVSKREGSEVVRRLRRYLDSGL';
     'MRTPMLLALLALAT--LCLAGRADAKPGDAESGK-GAAFVSKQEGSEVVKRLRRYLDHWL';
     'MRALTLLALLALAT--LCITGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL';
     'MRALTLLALLALAA--LCIAGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL';
     'MRALTLLALLALAA--LCIAGQAGAKPSGAESSK-GAAFVSKQEGSEVVKRPRRYLYQWL';]

mod_seq = seq(:, sum(seq == '-') < size(seq, 1)/2)

sum(seq == '-') gives you straight away the count of '-' per column.

[Rasif Ajwad]

Thanks. It worked. :)

But can you please explain this command as I am not familiar with it.