pheatmap in R - Error: logical subscript contains NAs

Question

[Rob]

I want to plot heatmap with the DESeq2 result. the heatmap will be for differentially expressed genes with FDR<0.05. but the data coming from DESeq2 includes NA which create problem when sub setting them for heatmap.

how can I deal with these NAs? this is the code:

## Differential abundance
alpha <- 0.05 #set the cutoff value
    ## subset only significant genes
    sig <- res[res$padj < alpha, ] 
    sig_genes <- rownames(sig)
    subset <- rdata[sig_genes,]
    ## log transform data for visualization
    tdata <- log2(subset + 0.5)
    mat <- as(tdata, "matrix")

    ## Set colours
    my_colour = list(
            Group = c("0" = "blue", "1" = "yellow"))

    #################
    pheatmap(symbreaks = FALSE,cluster_cols = FALSE,
             cluster_rows = TRUE,color = colorRampPalette(c("#f71616","#f71616","white",
    "#1919d4", "#1919d4"))(100),annotation_col = sample_org,
    annotation_colors = my_colour, mat, scale = "row")

Error is created here :

 sig <- res[res$padj < alpha, ]

error:

Error: logical subscript contains NAs

[newbio17]

Have you tried dropping the rows with NAs as p-values with na.omit(res) before applying the filter?

[Rob]

Hi newbio17 Thanks I tried na.omit(res)
this doesnot work my code was like then I get error again:

## Create metadata 

    sample_org <- data.frame(row.names = colnames(rdata), c(rep("0", 66), rep("1", 93)))
    colnames(sample_org) <- c("Group")

    dds <- DESeqDataSetFromMatrix(countData = rdata,
                                  colData = sample_org,
                                  design = ~Group)

    dd <- DESeq(dds)
    dd2 <- na.omit(dd)
    alpha <- 0.05

    res <- results(dd2)
    #Saving results
    write.csv(res,"res.csv")

    ## subset only significant genes
    sig <- res[res$padj < alpha,]
    sig_genes <- rownames(sig)
    subset <- rdata[sig_genes,]

error message is: in this line Im get error:

sig <- res[res$padj < alpha,]

Error: logical subscript contains NAs

Answer 1

[ATpoint]

You have to remove NAs first from the results as these were produced by the independent filtering. Spending time on R basics is recommended. All these functions like is.na() do not directly alter the object you give it. They produce logical lists (TRUE/FALSE) when the argument they test (in this case whether each entry is NA or not) are met or not. This you can then use to filter the original object. The ! below reverses the logical list, reads as "remove from res all those rows where the column padj is NOT NA):

res <- res[!is.na(res$padj),]