I would like to ask you one specific question regarding the DE analysis on an RNASeq dataset of samples, spanning a multi-factor experimental design. Briefly, unstimulated neutrophils of 4 healthy donors, were cultivated with distinct treatment conditions-that is, supernatant of organoids from different cancer/normal patient samples; There are also an additional treatment “sub-conditions”, which are not from an organoid patient sample, rather from 2 different "protein cocktails", inducing an”anti-tumoral” and “pro-tumoral” effect; A small subset of the relative phenotypic information is depicted below:
head(nets.pheno.dt.2,15) Donor_ID Condition Treatment CRC1_D10 Donor_10 Cancer CRC1 CRC1_D11 Donor_31 Cancer CRC1 CRC1_D5 Donor_5 Cancer CRC1 CRC2_D10 Donor_10 Cancer CRC2 CRC2_D5 Donor_5 Cancer CRC2 CRC2_D7 Donor_7 Cancer CRC2 CRC3_D10 Donor_10 Cancer CRC3 CRC3_D11 Donor_11 Cancer CRC3 CRC3_D5 Donor_5 Cancer CRC3 CRC4_D11 Donor_11 Cancer CRC4 CRC4_D5 Donor_5 Cancer CRC4 CRC4_D7 Donor_7 Cancer CRC4 CRC5_D10 Donor_10 Cancer CRC5 CRC5_D11 Donor_11 Cancer CRC5 CRC5_D5 Donor_5 Cancer CRC5 ...
In addition, after small processing and filtering, I created the following MDS plot:
y <- DGEList(counts=mat.counts, samples = nets.pheno.dt.2) y$genes <- data.frame(ENSEMBL=rownames(y), SYMBOL=final.annot$GENENAME) cond.NET <- y$samples$Condition keep.exprs <- filterByExpr(y, group=cond.NET) y.filt <- y[keep.exprs,, keep.lib.sizes=FALSE] dge <- calcNormFactors(y.filt, method = "TMM") lcpm <- cpm(dge, log=TRUE) plotMDS(lcpm, labels=dge$samples$Condition, col=as.numeric(dge$samples$Donor_ID))
MDS plot: labeling is based on the Condition factor, color on the 4 distinct donor IDs
Our major biological questions are: to have a general comparison of cancer vs normal “treatment” conditions; In addition if doable, to compare also the different “cancer-patient” treatment/condition effects, regardless of donor influence (i.e. CRC1 vs CRC2, Anti-tumoral versus tumoral, etc.);
However, as you can see from the MDS plot, where the different color denotes the different donor, clearly shows a dominant effect of the distinct donors, where the sub-conditions are barely separated, except mostly the "anti-tumoral" treatment; Hence, no differences between the "cancer" and "normal" treatment conditions; In addition, not all levels of experimental factors are represented in all “donors”; for example, donor 5 (dark blue color) does not include any “normal” treatment conditions;
Any suggestions or ideas on this matter would be greatly appreciated !!