RNASeq MATLAB: differenze tra le versioni

%% lncRNAs in prostate cancer in relation to Gleason score

%%% Read Input Data

path_images_folder = "MATLAB/RNASeq/images/";

% Loading read counts data

filenameGenData = "matrix_seminario.csv";

genData = readtable(filenameGenData, 'ReadRowNames', 1, 'ReadVariableNames', 1);

% Loading clinical data

filenameClinicalData = "design_gleason.csv";

clinicalData = readtable(filenameClinicalData);

% Count Features and Samples

nFeatures = size(genData, 1);

nPatients = size(genData, 2);

fprintf("There are %d features and %d samples\n", nFeatures, nPatients);

%%% Labels

% Create label vectors for high/low gleason score patients

labels_low = [];

labels_high = [];

for i = 1:nPatients

    label = clinicalData{i,3}{1};

    if label == "low"

        labels_low = [labels_low, i];

    else

        labels_high = [labels_high, i];

    end

end

nLows = numel(labels_low);

nHighs = numel(labels_high);

fprintf("There are:\n%d patients with low gleason\n%d patients with high gleason\n", nLows, nHighs);

=== Dimensionality Reduction ===

==== PCA ====

<syntaxhighlight lang="matlab" line>

%% PCA

genDataArray = table2array(genData);

genDataArrayT = genDataArray';

genDataArrayNormT = log2(genDataArrayT + 1);

[coeff,score,latent,tsquared,explained,mu] = pca(genDataArrayNormT, 'Centered', 0);

%%% Plot PC1 + PC2

figure

scatter(coeff(labels_low,1), coeff(labels_low,2), 1, 'black');

axis([-0.5 0.5 -0.5 0.5])

hold on 

scatter(coeff(labels_high,1), coeff(labels_high,2), 1, 'red');

%% tSNE

nDim = 3;

genDataEmbedding = tsne(genDataArrayT, 'Algorithm', 'exact', 'NumDimensions', nDim);

figure

if nDim == 2

    scatter(genDataEmbedding(labels_low,1), genDataEmbedding(labels_low,2), 3, 'black');

    hold on

    scatter(genDataEmbedding(labels_high,1), genDataEmbedding(labels_high,2), 3, 'red');

elseif  nDim == 3

    scatter3(genDataEmbedding(labels_low,1), genDataEmbedding(labels_low,2), ...

             genDataEmbedding(labels_low,3), 3, 'black');

    hold on

    scatter3(genDataEmbedding(labels_high,1), genDataEmbedding(labels_high,2), ...

             genDataEmbedding(labels_high,3), 3, 'red');

end

%% Normalizing Read Counts

counts = genDataArray;

% estimate pseudo-reference with geometric mean row by row

pseudoRefSample = geomean(counts,2);

nz = pseudoRefSample > 0;

ratios = bsxfun(@rdivide,counts(nz,:),pseudoRefSample(nz));

sizeFactors = median(ratios,1);

% transform to common scale

normCounts = bsxfun(@rdivide,counts,sizeFactors);

normCounts(1:10,:);

%% Boxplots

figure;

subplot(2,1,1)

maboxplot(log2(counts(:,1:4)),'title','Raw Read Count','orientation','horizontal')

ylabel('sample');

xlabel('log2(counts)');

subplot(2,1,2)

maboxplot(log2(normCounts(:,1:4)),'title','Normalized Read Count','orientation','horizontal')

ylabel('sample');

xlabel('log2(counts)');

%% Computing Mean, Dispersion and Fold Change

% consider the mean

% For each feature, we compute the mean across all the patients with

% high/low gleason

meanHigh = mean(normCounts(:,labels_high),2);

meanLow = mean(normCounts(:,labels_low),2);

% consider the dispersion

dispHigh = std(normCounts(:,labels_high),0,2) ./ meanHigh;

dispLow = std(normCounts(:,labels_low),0,2) ./ meanLow;

% plot on a log-log scale

figure;

loglog(meanHigh,dispHigh,'r.');

hold on;

loglog(meanLow,dispLow,'b.');

xlabel('log2(Mean)');

ylabel('log2(Dispersion)');

legend('High','Low','Location','southwest');

% compute the mean and the log2FC

meanBase = (meanHigh + meanLow) / 2;

foldChange = meanHigh ./ meanLow;

log2FC = log2(foldChange);

%% MA Plot

% plot mean vs. fold change (MA plot)

mairplot(meanHigh, meanLow,'Type','MA','Plotonly',true);

set(get(gca,'Xlabel'),'String','mean of normalized counts')

set(get(gca,'Ylabel'),'String','log2(fold change)')

%% Interactive MA Plot

mairplot(meanHigh,meanLow,'Labels',genData.Properties.RowNames,'Type','MA');

%% Create Gene Table

% create table with statistics about each gene

geneTable = table(meanBase,meanHigh,meanLow,foldChange,log2FC);

geneTable.Properties.RowNames = genData.Properties.RowNames;

% summary

summary(geneTable)

% preview

geneTable(1:10,:)

%% Constant Variance Link  

tConstant = nbintest(counts(:,labels_low),counts(:,labels_high),...

                  'VarianceLink','Constant');

h = plotVarianceLink(tConstant, 'Compare', 0);

% set custom title

h(1).Title.String = 'Variance Link on Low Samples';

h(2).Title.String = 'Variance Link on High Samples';

% save

saveas(h(1),path_images_folder + "tConstant_treated_0_prostate.png");

saveas(h(2),path_images_folder + "tConstant_untreated_0_prostate.png");

%% Local Regression Variance Link

tLocal = nbintest(counts(:,labels_low),counts(:,labels_high),'VarianceLink','LocalRegression');

h = plotVarianceLink(tLocal, 'Compare', 1);

% set custom title

h(1).Title.String = 'Variance Link on Low Samples';

h(2).Title.String = 'Variance Link on High Samples';

% save

saveas(h(1),path_images_folder + "tLocal_treated_prostate.png");

saveas(h(2),path_images_folder + "tLocal_untreated_prostate.png");

%% Histogram of P-values

h = figure;

histogram(tLocal.pValue,100)

xlabel('P-value');

ylabel('Frequency');

title('P-value enrichment')

saveas(h, path_images_folder + "p-values_histogram_gleason.png");

%%% Histogram of P-values without low count genes

h = figure;

lowCountThreshold = 10;

lowCountGenes = all(counts < lowCountThreshold, 2);

histogram(tLocal.pValue(~lowCountGenes),100)

xlabel('P-value');

ylabel('Frequency');

title('P-value enrichment without low count genes')

saveas(h, path_images_folder + "p-values_histogram_without_low_count_genes_gleason.png");

%% Multiple testing adjusted P-values

% compute the adjusted P-values (BH correction)

padj = mafdr(tLocal.pValue,'BHFDR',true);

% add to the existing table

geneTable.pvalue = tLocal.pValue;

geneTable.padj = padj;

% create a table with significant genes using adjusted p-values

sig = geneTable.padj < 0.1;

geneTableSig = geneTable(sig,:);

geneTableSig = sortrows(geneTableSig,'padj');

numberSigGenes = size(geneTableSig,1);

fprintf("There are %d significant genes on %d\n", numberSigGenes, nFeatures);

%% Identifying the Most Up-regulated and Down-regulated Genes

% find up-regulated genes

up = geneTableSig.log2FC > 1;

upGenes = sortrows(geneTableSig(up,:),'log2FC','descend');

numberSigGenesUp = sum(up);

fprintf('There are %d Up-regulated genes\n', numberSigGenesUp);

% find down-regulated genes

down = geneTableSig.log2FC < -1;

downGenes = sortrows(geneTableSig(down,:),'log2FC','ascend');

numberSigGenesDown = sum(down);

fprintf('There are %d Down-regulated genes\n', numberSigGenesUp);

%% Plot FC vs Mean

% A good visualization of the gene expressions and their significance is given by 

% plotting the fold change versus the mean in log scale and coloring the data points according 

% to the adjusted P-values.

figure

scatter(log2(geneTable.meanBase),geneTable.log2FC,20,geneTable.padj,'o')

colormap(flipud(cool(256)));

colorbar;

ylabel('log2(Fold Change)');

xlabel('log2(Mean of normalized counts)');

title('Fold change by FDR')

% You can see here that for weakly expressed genes (i.e. those with low means), 

% the FDR is generally high because low read counts are dominated by Poisson noise

% and consequently any biological variability is drowned in the uncertainties from the read counting.