3

我有一个包含基因信息的文本文件，例如基因之间的关系和部分关系。

此文本文件包含每个 GOTerm 的段落（GO 术语是一个包含特定代码编号的节点，例如：GO:0030436），它具有：Go 术语 ID（每个段落的第一行）和 isa（如果有）（以 isa 开头并以 isa 结尾）和 partof Go 术语（如果有）（以 partof: 开头并以 partof 结尾）来自此文本文件的一个小样本是：

GO:0030436
isa:
GO:0034297
GO:0043936
GO:0048315
end of isa
partof:
GO:0042243
end of partof
genes:
end of genes
GO:0034297
isa:
end of isa
partof:
end of partof
genes:
end of genes
GO:0043936
isa:
GO:0001410
GO:0034300
GO:0034301
GO:0034302
GO:0034303
GO:0034304
end of isa
partof:
end of partof
genes:
end of genes

我需要阅读这个文本文件并从中获取三个数据并制作一个包含 3 列的单元矩阵，如下所示：

map=

ID GoTerms               is_a            partof
GO:0030436              GO:0034297       GO:0042243
GO:0030436              GO:0043936              0
GO:0030436              GO:0048315              0
GO:0034297                   0                  0
GO:0043936              GO:0001410              0
GO:0043936              GO:0034300              0
GO:0043936              GO:0034301              0
GO:0043936              GO:0034302              0
GO:0043936              GO:0034303              0
GO:0043936              GO:0034304              0

请注意，如果每个 Go 术语包含多个术语或部分术语，我应该重复 Go 术语 ID 以使单元矩阵适合且组织良好。

关于如何制作这段代码的任何想法？

我尝试编写代码，但它不起作用，因为我不知道如何使用超过 1 个 isa 和部分术语：

s={};
        fid = fopen('Opt.pad'); % read from the certain text file 
        tline = fgetl(fid); 
        while ischar(tline) 
           s=[s;tline]; 
           tline = fgetl(fid); 
        end 
% find start and end positions of every [Term] marker in s 
    terms = [find(~cellfun('isempty', regexp(s, '\GO:\w*'))); numel(s)+1];
      % for every [Term] section, run the previously implemented regexps
      % and save the results into a map - a cell array with 3 columns      map = cell(0,3);
      for term=1:numel(terms)-1
          % extract single [Term]  data
          s_term = s(terms(term):terms(term+1)-1);          % match regexps
          %To generate the GO_Terms vector from the text file
          tok = regexp(s_term, '^(GO:\w*)', 'tokens');
          idx = ~cellfun('isempty', tok); 
          GO_Terms=cellfun(@(x)x{1}, (tok(idx)));          %To generate the is_a relations vector from the text file
          tok = regexp(s_term, '^isa: (GO:\w*)', 'tokens'); 
          idx = ~cellfun('isempty', tok); 
          is_a_relations  =cellfun(@(x)x{1}, (tok(idx)));          %To generate the part_of relaions vector from the text file
          tok = regexp(s_term, '^partof: (GO:\w*)', 'tokens'); 
          idx = ~cellfun('isempty', tok); 
          part_of_relations =cellfun(@(x)x{1}, (tok(idx)));          % map. note the end+1 - here we create a new map row. Only once!
          map{end+1,1} = GO_Terms;
          map{end,  2} = is_a_relations;
          map{end,  3} = part_of_relations;
      end      map( cellfun(@isempty, map) ) = {0};
4

1 回答 1

0

A short and simple solution (albeit perhaps not the fastest):

% # Parse text file
C = textread('Opt.pad', '%s', 'delimiter', '');

% # Obtain indices for isa elements
idx = reshape(find(~cellfun(@isempty, strfind(C, 'isa')))', 2, []);
isa = arrayfun(@(x, y)x + 1:y - 1, idx(1, :), idx(2, :), 'Uniform', false);

% # Obtain indices for partof elements
idx = reshape(find(~cellfun(@isempty, strfind(C, 'partof')))', 2, []);
partof = arrayfun(@(x, y)x + 1:y - 1, idx(1, :), idx(2, :), 'Uniform', false);

% # Obtain indices of GO term elements and IDs
go = find(cellfun(@(s)any(strfind(s, 'GO:')), C));
id = go(~ismember(go, [isa{:}, partof{:}]));

% # Construct a new cell array
N = cellfun(@(x, y)max([numel(x), numel(y), 1]), isa, partof);
k = cumsum([1, N(1:end - 1)]);
X = cell(sum(N), 3); % # Preallocate memory!
repcell = @(x, n)arrayfun(@(y)x, 1:n, 'Uniform', false);
for ii = 1:numel(id)
    idx = k(ii):k(ii) + N(ii) - 1;
    X(idx, 1) = repcell(C{id(ii)}, N(ii));
    X(idx, 2) = [C{isa{ii}}, repcell('0', N(ii) - numel(isa{ii}))];
    X(idx, 3) = [C{partof{ii}}, repcell('0', N(ii) - numel(partof{ii}))];
end

This should yield the following output:

X = 

    'GO:0030436'    'GO:0034297'    'GO:0042243'
    'GO:0030436'    'GO:0043936'    '0'         
    'GO:0030436'    'GO:0048315'    '0'         
    'GO:0034297'    '0'             '0'         
    'GO:0043936'    'GO:0001410'    '0'         
    'GO:0043936'    'GO:0034300'    '0'         
    'GO:0043936'    'GO:0034301'    '0'         
    'GO:0043936'    'GO:0034302'    '0'         
    'GO:0043936'    'GO:0034303'    '0'         
    'GO:0043936'    'GO:0034304'    '0'
于 2012-12-16T18:40:57.583 回答