how to replace characters into digits

2 次查看（过去 30 天）
Sharen H
Sharen H 2013-2-27
i have to replace the each characters using the following digits s=ACGT
i have to replace as
'A' then 11
'C' then 00
'G' then 01
'T' then 10
  1 个评论
Jan
Jan 2013-2-27
You write 11 without surrounding quotes. This isn't an accident, correctly? What do you want as output? How long is the input?

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Azzi Abdelmalek
Azzi Abdelmalek 2013-2-27
编辑：Azzi Abdelmalek 2013-2-27
clear
s='ACGT'
e=['11';'00';'01';'10']
in='AGCTAG' % Your initial data
out=in
for k=1:numel(s)
out=regexprep(out,s(k),e(k,:))
end
  3 个评论
显示 1更早的评论
Jan
Jan 2013-2-28
STRREP is less powerful, but much faster than REGEXPREP.
Cedric
Cedric 2013-2-28
STRREP is probably the best solution. It is a little slower than my solution, but more memory friendly.

请先登录，再进行评论。

更多回答（3 个）

Jan
Jan 2013-2-28
And a lookup table:
seqIn = 'ACGTTGCA'
table = repmat('0', 2, 255);
table(1, 'AT') = '1';
table(2, 'AG') = '1';
result = reshape(table(:, seqIn), 1, []);
Does this work? I do not have access to Matlab currently.
  3 个评论
显示 1更早的评论
Cedric
Cedric 2013-2-28
编辑：Cedric 2013-2-28
Actually STRREP still wins up to me. The solution based on ISMEMBER is profiled between REGEXP and the two solutions based on array indexing. But the latter require much more memory during the indexing operation than STRREP. You start seeing that with ~1e8 chars. On my laptop with 8GB RAM for example, only STRREP can treat more than 2e8 chars without swapping (or being killed by Process Lasso).
Jan
Jan 2013-2-28
@Azzi: Is this a typo?! Your function needs 14 secs with REGEXPREP and 0.007 secs with STRREP? Then my minor suggestion caused a speedup of a factor 1900? Wow, this would be the most efficient suggestion I ever gave. And it would be a strong hint to warn for the low efficiency of regexprep in this forum.

请先登录，再进行评论。

Cedric
Cedric 2013-2-27
编辑：Cedric 2013-2-27
If you need to process long sequences, you might want to optimize a little the efficiency.. a MEX-based solution would be most efficient I guess, but here is one way you could go using basic MATLAB only..
If you want to replace character 'A' with a numeric array (1,1) and so on, you can do:
aa = 'ACGT' ;
seq = 'AAGCTCAGGTTCA' ;
rep = zeros(2, max(aa), 'uint8') ;
rep(:,aa) = [1 0 0 1; 1 0 1 0] ;
result = reshape(rep(:,seq), 1, []) ;
This outputs the numeric array:
result =
1 1 1 1 0 1 0 0 1 0 0 0 1 1 0 1 0 1 1 0 1 0 0 0 1 1
If you want to replace character 'A' with characters '11' and so on, you can do:
aa = 'ACGT' ;
seq = 'AAGCTCAGGTTCA' ;
rep = zeros(2, max(aa), 'uint8') ;
rep(:,aa) = ['11'; '00'; '01'; '10'].' ;
result = char(reshape(rep(:,seq), 1, [])) ;
This outputs the string '11110100100011010110100011'.
EDIT: note that there are slightly different ways of doing this a little slower but with a more memory-friendly approach.
Cheers,
Cedric
Jos (10584)
Jos (10584) 2013-2-28
Here is a simpler approach than looping over REGEXPREP or STRREP:
seqIn = 'ACGTTGCA' % input sequence
letters = 'ACGT' ;
symb = {'11','00','10','01'} ; % stored as a cell array of strings!
[tf,idx] = ismember(seqIn,letters) ;
seqOut = [symb{idx}]

类别

在 Help Center 和 File Exchange 中查找有关 Migrate GUIDE Apps 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by