Microarray data analysis

Question

[newyaso]

Can any one help me perform clustering and gene enrichment analysis.

I had discovered my differential gene expression list using these commands

targets <- readTargets("targets.txt")
f <- function(x) as.numeric(x$Flags > -50)
RG <- read.maimages(targets,source="genepix", columns=list(R="F635 Median",G="F532 Median"),wt.fun=f)
RG$genes <- readGAL()
RG.b <- backgroundCorrect(RG, method="normexp", offset=50)
MA <- normalizeWithinArrays(RG.b)
biolrep <- c(1, 1, 2, 2)
design <- c(1, -1, 1, -1)
library (statmod)
corfit <- duplicateCorrelation(MA, design, ndups = 1, block = biolrep)
fit <- lmFit(MA, design, block = biolrep, cor = corfit$consensus)
fit <- eBayes(fit)
fit2 <- topTable(fit,  number=600, adjust = "BH", p.value=0.05, lfc=2)
head(fit2)
      Block Column Row     Name    ID                               Loc                                                                GO
2729     48     15  31    Blank Blank                             Empty                                                               N/A
19312    47      6  19 TR046495   G13                    chr04:27400414                                                               N/A
2659     47     28  15    BLANK   K13                             Empty                                                               N/A
15465    45     16  29 TR042648   E13                  LOC_Os03g50030.1                                                        GO:0016787
42447    33     17   6 TR069630   E16                  LOC_Os11g42100.1                                                        GO:0008150
43985    14      6  12 TR071168   F09 LOC_Os03g50010.1;LOC_Os03g50010.2 GO:0005739;GO:0006950;GO:0016787;GO:0005739;GO:0006950;GO:0016787
                                         Annotation     logFC  AveExpr         t      P.Value  adj.P.Val        B
2729                                            N/A -3.615243 12.01787 -15.33867 7.085888e-07 0.02098715 5.965012
19312                                           N/A -2.925755 10.49668 -12.84876 2.517883e-06 0.02098715 5.070669
2659                                            N/A -3.027424 10.75164 -12.52566 3.017522e-06 0.02098715 4.933931
15465         Phospholipase A2, putative, expressed  3.294050 11.45860  11.80853 4.581567e-06 0.02098715 4.610336
42447 Leucine Rich Repeat family protein, expressed -2.729774 10.97718 -11.75398 4.733877e-06 0.02098715 4.584526
43985                    Toc64, putative, expressed  3.377292 11.35714  11.71446 4.847810e-06 0.02098715 4.565715

So please I want to cluster my data and draw heatmap then perform gene enrichment analysis

Note This data was a time course experiment

Answer 1

[Charles Warden]

In general, I would use heatmap.2 for heatmaps (using the normalized expression values for the differentially expressed probes) and web-based tools (like GATHER, DAVID, FuncAssociate, etc.) for functional enrichment (GO, KEGG, etc).

However, for the functional enrichment, you will need some sort of standard nomenclature (typically official gene symbols), and I don't see anything like that I recognize. You could try to figure this out, but it looks like you already have GO mappings in your annotations. So, if it was me, I'd probably manually count the relevant statistics to do the hypergeometric or Fisher Exact test manually.

As for the heatmap, I'm not sure what is the most efficient strategy to get a heatmap from your limma result, but this is what I found with a quick Google search:

http://www2.warwick.ac.uk/fac/sci/moac/people/students/peter_cock/r/heatmap/