Question

RNAseq and heatmap

0

Entering edit mode

Merlin ▴ 10

@merlin-15723

Last seen 4.3 years ago

Vancouver

Hi guys,

When at the end of the analysis I export the results with this code:

write.csv(as.data.frame(resOrdered), 
          file="condition_treated_results.csv")

I have a file containing a list of genes with Log2FC and p-value. let say I have two conditions treated and untreated (6 and 2 ), are the values of gene expression of two replicates (untreated) normalized and compared with the values of the 6 replicates (treated) ? I mean are this taken as a single condition isn't it?

What if I have more conditions? let say 4 different conditions, to create that table 3 will be considered as a unique and one as reference condition?

Thank you

also, to make a heatmap where all the biological replicates cluster together I could take the top 1000 gene that are similar among replicates and different with the reference level but I don't know if that is right and/or DESeq2 can do that, what do you guys think?

Thanks

deseq2 RNA-seq Heatmap • 1.6k views

ADD COMMENT • link updated 4.8 years ago by Michael Love 41k • written 4.8 years ago by Merlin ▴ 10

score 2 · Answer 1 · 2019-07-24

2

Entering edit mode

Michael Love 41k

@mikelove

Last seen 1 hour ago

United States

Take a look at the workflow which describes what shows up in the results table:

https://bioconductor.org/packages/rnaseqGene

To contrast multiple groups, you need to call results() multiple times using the contrast argument. See:

?results

For the last case, you can find these genes by performing a likelihood ratio test. If you have a dds with design ~condition, you can do:

dds <- DESeq(dds, test="LRT", reduced=~1)
res <- results(dds)
heatmap.genes <- head(order(res$padj), n)

where n is the number of genes for the heatmap.

ADD COMMENT • link 4.8 years ago Michael Love 41k

0

Entering edit mode

Thank you, now the first part is clear enough,

For the heatmap,I tried this code as you say but I have this result,

pheatmap(heatmap.genes) Error in hclust(d, method = method) : must have n >= 2 objects to cluster

I think is missing the information of the samples name which in my previous command for heatmao was rld,

topGenes<- head(order(rowVars(assay(rld)), decreasing = TRUE),1000)

Can you tell me how to fix it?

I used this to show the heatmap

library("pheatmap") pheatmap(heatmap.genes)

thank you

ADD REPLY • link 4.8 years ago Merlin ▴ 10

0

Entering edit mode

Looks like you’re having some issues going from the transformer data to the heatmap. I’d recommend go one step at a time and make sure you have what you are expecting at each step. Eg heatmap.genes is not data, it’s a set of numeric indices. You should get in the habit of looking at the various objects to see what they represent, using eg head() and dim(), etc.

ADD REPLY • link 4.8 years ago Michael Love 41k