B_spinePreference.m

% Fig. 2B: Innervation preference for spine and shaft of axons targeting apical
% dendrites

% Author: Ali Karimi <ali.karimi@brain.mpg.de>

%% set-up
util.clearAll;
c = util.plot.getColors();
outputDir = fullfile(util.dir.getFig(2),'B');
util.mkdir(outputDir);

%% Write excel sheet of result (independent function)
dendrite.synIdentity.writeSpinePreferenceExcelSheet;

%% Get the ratio of each postsynaptic type
synRatio = dendrite.synIdentity.getSynapseMeasure('getSynRatio');
synRatio.L2{'L5A','Spine'}{1} = [];
synRatio.L1{'layer5AApicalDendriteSeeded','Spine'}{1} = [];

%% Plot the fraction of single spine innervation
layers = fieldnames(synRatio);
seedTypes = synRatio.L1.Properties.VariableNames;
colors = {util.plot.getColors().l2vsl3vsl5([1:3,5]),...
    {c.l2color,c.dlcolor,c.l5Acolor}};
ylimitsKS = [0,15];
for l = 1:2
    layerRatios = synRatio.(layers{l});
    curColor = colors{l};
    for sType = 1:2
        fname = strjoin({layers{l},[seedTypes{sType},'.svg']},'_');
        fh = figure('Name',fname); ax = gca;
        hold on
        for cType = 1:height(layerRatios)
            thisCellRatios = layerRatios{cType,sType}{1};
            if ~isempty(thisCellRatios)
                % Plot histogram/KSdensity
                util.plot.histogramAndKsdensity...
                    (thisCellRatios.Spine,curColor{cType},ylimitsKS);
            end
        end
        % Save
        util.plot.cosmeticsSave(fh,ax,1.8,1.4,outputDir,fname);
        hold off
    end
end

%% For text: number of axons in each group
for l = 1:2
    curRatio = synRatio.(layers{l});
    tableOFNumberOFAxons.(layers{l}) = array2table(cellfun(@util.table.height,...
        curRatio.Variables));
    tableOFNumberOFAxons.(layers{l}) = util.table.copyRVNames...
        (curRatio,tableOFNumberOFAxons.(layers{l}));
    fprintf('For text, number of axons in %s datasets, per group:\n',layers{l})
    disp(tableOFNumberOFAxons.(layers{l}));
end

%% Get synapse number range and misclassification rate
% Get range of axonal input number
synCount = dendrite.synIdentity.getSynapseMeasure('getTotalSynNumber');
% Remove duplicate of L5tt in L5st spine annotations
synCount.L1.Spine{end} = [];
synCount.L2.Spine{end} = [];
allSynCount = cat(1,synCount.L1.Spine{:},synCount.L2.Spine{:},...
    synCount.L1.Shaft{:},synCount.L2.Shaft{:});
assert(height(allSynCount) == 430);
disp(['Total synapse range per axon:', ...
    num2str([min(allSynCount.totalSynapseNumber),...
    max(allSynCount.totalSynapseNumber)])]);
% Current total misclassification rate (checked with excel sheets): 6.28%
% see below
% Load axon switching fraction
axonSwitchFraction = dendrite.synIdentity.loadSwitchFraction;
total = zeros(2,1);
misclassified = zeros(2,1);
for l = 1:2
    curNum = tableOFNumberOFAxons.(layers{l});
    curCorrection = axonSwitchFraction.(layers{l});
    % Shaft and spine columns switched along for number and fractions
    curMisclassified = curNum.Variables .* curCorrection.Variables;
    total(l) = sum(curNum.Variables,'all');
    misclassified(l) = sum(curMisclassified,'all');
end
% Total number of synapses per layer
disp ('Total number of synapses with axons investigated:')
disp(['L1: ', num2str(total(1))]);
disp(['L2: ', num2str(total(2))]);

% Percent of misclassified axons
percentMisclassified = (sum(misclassified)./sum(total))*100;
disp (['Percent misclassified: ',num2str(percentMisclassified)])
%% L2 and DL 90% and 80% minimum fraction percentiles in L2

seedTypes = {'Spine','Shaft'};
adTypes = {'L2','DL'};
for seedT = 1:2
    for adT = 1:2
        curRatio = synRatio.L2.(seedTypes{seedT}){adT}.(seedTypes{seedT});
        disp([seedTypes{seedT},'_seeded_',adTypes{adT}]);
        disp(['TotalNum: ',num2str(length(curRatio))]);
        disp(['At least 90% on seed type:',num2str(sum(curRatio >= .9))]);
        disp(['At least 80% on seed type:',num2str(sum(curRatio >= .8))]);
    end
end