Applying high-dimensional single-cell

Tumoursareacomplexmixtureofmalignant,immuneandstromalcells,whichoftenhavesubstantiallevelsofintratumourandintertumourheterogeneity.Thetumourmicroenvironment(TME)comprisesanamalgamoftumour-promotingandanti-tumoursignalsthatareabletomodulatetumourgrowthandinfluencetumourevolution.Bulkgenomicandtranscriptomicanalyseshaveprovidedvaluableinsightsintotheseprocesses,buttheaveragingofsignalsfromlargenumbersofcellsrenderedbythesemethodsoftenobscuresspecificsubpopulationsand/orcellularstates.Indeed,rareanduniquecellularsubtypesorstatesthatmightbeinstrumentalindiseasebiology,suchaspotential‘cancerstemcells’orimmunecellscrucialforatherapeuticresponse,mightnotbedetectedinbulkanalysesderivedbyaveragingdatafrommillionsofcells.Forcancerimmunotherapy,andthislimitationreducesthepowertodistinguishthediversityofcellularsubtypesandstatespresentintheTME.Breakthroughsinsequencingtechnologies,cell-isolatingmicrofluidicsandanalyticalbioinformaticshaveledtoarapidproliferationofsingle-cellanalysismethodologies,andhavebeenreviewedextensivelyelsewhere3–6.Thesetechniquesenablethedetectionofmanyanalytesatsingle-cellresolutioninalargenumberofcells(initiallyafewhundred,butnowthousandstohundredsofthousandsofcells)inanincreasinglycost-effectivemanner.Theconcomitantblossomingoftheavailabilityofbioinformatictoolshasenabledresearcherstointerrogatethemassivedatasetsprovidedbysuchmethodsinordertoexploitthecovariatepowerattheheartofsingle-celldata,visualizethediversityofcellularsubtypesandelucidatetheinteractionsbe