How to make Heatmap about DEG?

Question

[arabi]

I've started learning RNA-seq only recently. I want to compare the DEG expression levels in WT(Wild Type) and three mutants(mutantA, mutantB, mutantC ; lacked a certain genes).

I tried to use hisat2, stringtie and ballgown(refer to https://www.nature.com/articles/nprot.2016.095). I've (probably) done these, but I don't know what to do next...

First of all, How to get DEG(diffentially expressed gene)?

I think that :

• compare WT with mutantA

• compare WT with mutantB

• compare WT with mutantC

  ↓

  Merge there 3 results

I used TCC-GUI and carried out the above ,but I failed.

Will you kindly tell me how to get DEG and flow of making heatmap?

Thanks in advance for your help.

[ATpoint]

Please read any of the RNA-seq workflows and follow the example code, e.g. https://www.bioconductor.org/packages/devel/workflows/vignettes/rnaseqGene/inst/doc/rnaseqGene.html I personally do not prefer the hisat/stringtie workflow as assembling the transcriptome via stringtie is simply not necessary in most cases and only complicates the analysis. The linked workflow suggests a good workflow that is simpler to follow. Also read a manual towards heatmaps, e.g. the ComplexHeatmap package. There are also plenty of related questions here at Biostars towards both topics. Please read them.

[arabi]

Oh... I'll check the link.Thank you for answering my question! Sorry, I couldn't understand what you said completely.

as assembling the transcriptome via stringtie is simply not necessary in most cases and only complicates the analysis.

Sorry for inconveniencing you but would you explain a little more in detail?

[ATpoint]

stringtie is a tool to assemble a transcriptome. That means it takes the reads from RNA-seq (alignments) and tries to put them together into transcripts followed by quantifying reads against that transcriptome. This is in most cases not necessary for organisms wehre well-curated transcriptome assemblies already exist such as human, mouse, other model organisms. Therefore I do not see the point in using it. Also this ballgown tool is pretty clunky to me. It still uses the deprecated FPKM normalization which regularily fails to perform well in benchmarking studies. I personally find the manual rather unintuitive, especially compared to the excellent alternatives for gene-level DEG analysis such as DESeq2, limma and edgeR. I agree though that as a beginner it is difficult to decide for a pipeline since this workflow you follow is published high from reputable authors.

Still, I suggest you follow the workflow that I linked. It is well-maintained, authors are active at Bioconductor support forum and it has a large user base, plus it is (at least for me) more intuitive to use than this hisat/stringtie/ballgown workflow.

[arabi]

Thank you for your reply. Your explanations were clear and understandable! I'll try to use this workflow.