Hi friends,
I am an undergrad and was asked by a post doc to help analyze an Illumina Hi-seq single end RNA seq dataset. Needless to say I am a complete novice to this type of analysis and a little bit overwhelmed. We are working with S. cerevisiae; one wildtype condition and five test conditions (two biological replicates each, so 12 total FASTQ files). We are not interested in novel gene/splice variants and simply want to explore gene expression profiles of known genes. So I first aligned the reads to an annotated reference genome with TopHat then used CuffDiff using a reference transcriptome along with the BAM files from TopHat.
My cuffdiff input looked like:
$ cuffdiff -p 8 -o cuffdiff_out -b genome.fa -u genes.gtf -L WT,ies2,arp8,ino80,ies6,arp5 ./BY4741_1_accepted_hits.bam,./BY4741_2_accepted_hits.bam
./ies2_1_accepted_hits.bam,./ies2_2_accepted_hits.bam
./arp8_1_accepted_hits.bam,./arp8_2_accepted_hits.bam
./ino80_1_accepted_hits.bam,./ino80_2_accepted_hits.bam
./ies6_1_accepted_hits.bam,./ies6_2_accepted_hits.bam
./arp5_1_accepted_hits.bam,./arp5_2_accepted_hits.bam
My question is: How do I generate a heatmap showing the fold change in gene expression of each of the 5 mutants compared to WT using cummeRbund?
With cummerBund I was able to generate a heatmap with
> csHeatmap(myGenes, cluster="both") where 'myGenes' was a list of the top 50 differentially expressed genes. I have attached the figure.
However, I would rather have a heatmap where there is no WT (BY4741) column, and instead just shows fold change of the five mutant samples relative to WT. Is there a way to do this? If so, is there a way to specify the color scheme so that down-regulated genes are clearly one color (like blue) and up-regulated genes are clearly another color (like red), and genes that have no change are just black?
Again, I'm new to this kind of analysis, and completely new to R, but cummeRbund seems like the easiest/most powerful way to do this. Advice?
I am an undergrad and was asked by a post doc to help analyze an Illumina Hi-seq single end RNA seq dataset. Needless to say I am a complete novice to this type of analysis and a little bit overwhelmed. We are working with S. cerevisiae; one wildtype condition and five test conditions (two biological replicates each, so 12 total FASTQ files). We are not interested in novel gene/splice variants and simply want to explore gene expression profiles of known genes. So I first aligned the reads to an annotated reference genome with TopHat then used CuffDiff using a reference transcriptome along with the BAM files from TopHat.
My cuffdiff input looked like:
$ cuffdiff -p 8 -o cuffdiff_out -b genome.fa -u genes.gtf -L WT,ies2,arp8,ino80,ies6,arp5 ./BY4741_1_accepted_hits.bam,./BY4741_2_accepted_hits.bam
./ies2_1_accepted_hits.bam,./ies2_2_accepted_hits.bam
./arp8_1_accepted_hits.bam,./arp8_2_accepted_hits.bam
./ino80_1_accepted_hits.bam,./ino80_2_accepted_hits.bam
./ies6_1_accepted_hits.bam,./ies6_2_accepted_hits.bam
./arp5_1_accepted_hits.bam,./arp5_2_accepted_hits.bam
My question is: How do I generate a heatmap showing the fold change in gene expression of each of the 5 mutants compared to WT using cummeRbund?
With cummerBund I was able to generate a heatmap with
> csHeatmap(myGenes, cluster="both") where 'myGenes' was a list of the top 50 differentially expressed genes. I have attached the figure.
However, I would rather have a heatmap where there is no WT (BY4741) column, and instead just shows fold change of the five mutant samples relative to WT. Is there a way to do this? If so, is there a way to specify the color scheme so that down-regulated genes are clearly one color (like blue) and up-regulated genes are clearly another color (like red), and genes that have no change are just black?
Again, I'm new to this kind of analysis, and completely new to R, but cummeRbund seems like the easiest/most powerful way to do this. Advice?
Comment