groupSIFT_imputeMissingSubjects.m

% groupSIFT_imputeMissingSubjects()--Takes two groupSIFT-saved individual
%                                    subject data to perform missing data
%                                    imputation using MDI Toolbox 4. The
%                                    default algorithm used is Trimmed
%                                    Scores Regression (TSR). The output
%                                    variable may be saved by using
%                                    Matlab fucntion writetable().
%
% Usage
%       outputTable = groupSIFT_imputeMissingSubjects(xlsx1, xlsx2)
%
% Input
%       xlsx1: Fullpath for the *.xlsx file generated by groupSIFT that contains individual subject data for each edge.
%       xlsx2: Fullpath for the *.xlsx file generated by groupSIFT that contains individual subject data for each edge.
%
% Output
%       outputTable: Matlab table variable that contains corresponding pairs between xlsx1 and xlsx2 while missing variables imputed.
%
% History
% 06/05/2020 Makoto. Created.

% Copyright (C) 2020, Makoto Miyakoshi (mmiyakoshi@ucsd.edu) , SCCN,INC,UCSD
%
% Redistribution and use in source and binary forms, with or without
% modification, are permitted provided that the following conditions are met:
%
% 1. Redistributions of source code must retain the above copyright notice,
% this list of conditions and the following disclaimer.
%
% 2. Redistributions in binary form must reproduce the above copyright notice,
% this list of conditions and the following disclaimer in the documentation
% and/or other materials provided with the distribution.
%
% THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
% AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
% IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
% ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
% LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
% CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
% SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
% INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
% CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
% ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF
% THE POSSIBILITY OF SUCH DAMAGE.



function outputTable = groupSIFT_imputeMissingSubjects(xlsx1, xlsx2)

% Set path to the Missing Data Imputation toolbox.
pathCheck = which('pcambtsr');
if isempty(pathCheck)
    error('Set path to MDI Toolbox 4.')
end

% Load two datasets.
[NUM1,TXT1,RAW1] = xlsread(xlsx1);
[NUM2,TXT2,RAW2] = xlsread(xlsx2);

% Preliminary solution to export to excel.
dataSheet = cell(1,1);
dataSheet{1,1} = 'from';
dataSheet{2,1} = 'to';
dataSheet{3,1} = 'centroidTimeS';
dataSheet{4,1} = 'centroidFreqHz';
dataSheet{5,1} = 'clusterSize';

% Loop across edges.
for edgeIdx = 1:(size(RAW1,2)-1)/2
    
    % Find column indices for variables and subject names.
    currentEdgeVariableIdx = edgeIdx*2;
    currentEdgeSubjIdx     = currentEdgeVariableIdx+1;

    % Obtain the variables and subject names.
    currentEdgeVar1 = cell2mat(RAW1(7:end,currentEdgeVariableIdx));
    currentSubjVar1 = RAW1(7:end,currentEdgeSubjIdx);
    currentEdgeVar2 = cell2mat(RAW2(7:end,currentEdgeVariableIdx));
    currentSubjVar2 = RAW2(7:end,currentEdgeSubjIdx);    
    
    % Remove NaNs.
    nanMask1 = isnan(currentEdgeVar1);
    currentEdgeVar1(nanMask1) = [];
    currentSubjVar1(nanMask1) = [];
    nanMask2 = isnan(currentEdgeVar2);
    currentEdgeVar2(nanMask2) = [];
    currentSubjVar2(nanMask2) = [];
    
    % Extract common subject IDs.
    subj1 = cellfun(@(x) x(4:9), currentSubjVar1, 'uniformoutput', false);
    subj2 = cellfun(@(x) x(4:9), currentSubjVar2, 'uniformoutput', false);
    
    % Generate data matrix with missing values.
    allUniqueSubj = unique([subj1; subj2]);
    missingDataMatrix = nan(length(allUniqueSubj),2);
    [~,subj1Idx] = intersect(allUniqueSubj, subj1);
    [~,subj2Idx] = intersect(allUniqueSubj, subj2);
    missingDataMatrix(subj1Idx,1) = currentEdgeVar1;
    missingDataMatrix(subj2Idx,2) = currentEdgeVar2;

    % Perform missing value interpolation.
    imputedData = pcambtsr(missingDataMatrix, size(missingDataMatrix,2), 5000, 1e-10);
    
    % Build the new output.
    dataSheet(1:5, edgeIdx*3-1) = RAW1(2:6,currentEdgeVariableIdx);
    dataSheet(6:6+size(imputedData,1)-1, edgeIdx*3-1) = allUniqueSubj;
    dataSheet{4, edgeIdx*3}   = 'Data1';
    dataSheet{5, edgeIdx*3}   = ['Missing:', num2str(find(isnan(missingDataMatrix(:,1)')))];
    dataSheet(6:6+size(imputedData,1)-1, edgeIdx*3)   = num2cell(imputedData(:,1));
    dataSheet{4, edgeIdx*3+1} = 'Data2';
    dataSheet{5, edgeIdx*3+1} = ['Missing:', num2str(find(isnan(missingDataMatrix(:,2)')))];
    dataSheet(6:6+size(imputedData,1)-1, edgeIdx*3+1) = num2cell(imputedData(:,2));
end

% Convert cell to table.
outputTable = cell2table(dataSheet);