How to see if characters are present in a string array.

I am trying to write some code that will take a short amino acid sequence, ex. 'GSA' and then search through a string array of sequences to find the number and index of matches, but I would like it to ignore the order of the characters. As long as each character is present, I would like to consider it a hit.

Here is the code I have so far, which kind of works. InputSeq is the sequence I would like to search for, and AAseq is the string array of sequences that I would be searching through. This code only produces a match if all characters are present AND the order is correct.

InputSeq = "GSA";

AAseq = [ SGD; SGS; SGA; SGV; SGS; SGA; SGD; SGS; SGS; SGY; SGD; SGS; SGI.........];

result = ismember(InputSeq, AAseq)

This kind of works, but it will not register a match if the order of the characters does not match.

Answer 1

[Stephen23]

Assuming that all string elements contain exactly the same number of characters, then you can do this easily with basci logical operations on character arrays:

A = "GSA";
B = ["SGD";"SGS";"SGA";"SGV";"SGS";"SGA";"SGD";"SGS";"SGS";"SGY";"SGD";"SGS";"SGI"]

B = 13×1 string array
    "SGD"
    "SGS"
    "SGA"
    "SGV"
    "SGS"
    "SGA"
    "SGD"
    "SGS"
    "SGS"
    "SGY"
    "SGD"
    "SGS"
    "SGI"

X = all(sort(char(A))==sort(char(B),2),2)

X = 13×1 logical array
   0
   0
   1
   0
   0
   1
   0
   0
   0
   0

Or without sorting:

X = all(any(char(A)==permute(char(B),[1,3,2]),3),2)

X = 13×1 logical array
   0
   0
   1
   0
   0
   1
   0
   0
   0
   0

[Elijah Roberts]

Thanks! This worked the best for me, but I had to make some changes to the way I sorted my character array. They way you coded it, it alphabetized by row and didn't alphabetize the columns. I solved it by doing this. I think it is because my variable was a character array already rather than a string.

A = 'GSA'

B = ['SGD';'SGS';'SGA';'SGV';'SGS';'SGA';'SGD';'SGS';'SGS';'SGY';'SGD';'SGS';'SGI']

for i = 1:length(B)

B(i,:) = sort(B(i,:));

end

Result = all(sort(A) == B, 2);

MatchIdx = find(Result == 1);

MatchIdx =

3

6

[Stephen23]

You don't need the loop, youc an simply specify the sort dimension argument:

A = 'GSA'

A = 'GSA'

B = ['SGD';'SGS';'SGA';'SGV';'SGS';'SGA';'SGD';'SGS';'SGS';'SGY';'SGD';'SGS';'SGI']

B = 13×3 char array
    'SGD'
    'SGS'
    'SGA'
    'SGV'
    'SGS'
    'SGA'
    'SGD'
    'SGS'
    'SGS'
    'SGY'
    'SGD'
    'SGS'
    'SGI'

X = all(sort(A)==sort(B,2),2)

X = 13×1 logical array
   0
   0
   1
   0
   0
   1
   0
   0
   0
   0

[Elijah Roberts]

Yep, you're right! That worked. Thank you!

Answer 2

[Walter Roberson]

You could use multiple contains() tests.

But I suggest that instead you do something like

ismember(sort(char(InputSeq)), cellfun(@sort, cellstr(AAseq), 'uniform', 0))

[Elijah Roberts]

That is only returning true or false i.e. "InputSeq is found somewhere in AAseq." I would like to know get a logic array of the same size as AAseq, so I can get all of the indeces of the matching sequences.

I had some luck with this, I also trimmed the input sequence down to 'GS,' and the AAseq are all two characters long as well

Matches = ismember(InputSeq, AAseq); (both variables are char arrays)

This gave me a 96x2 logic array. Column one seems to be "is G a member" and column 2 is "is S a member"

This kind of works for me. If I can get the row indeces where both columns are true I will be good.

I tried this

MatchIndex = find(Matches == [1 1])

but it just gave me every index where there is a 1, rather than giving me indeces where both columns are 1.

[Walter Roberson]

ismember( cellfun(@sort, cellstr(AAseq), 'uniform', 0), sort(char(InputSeq)) )

You could also strcmp()