PCA plot for miRNA

Question

[struggler]

Hi everyone,

I have RNAseq and sRNAseq data for one specific condition in plant. RNAseq data shows very well behaved pattern on PCA plot while miRNA does not show that pattern (not even biological sense) on PCA. Could anyone please help me to understand this issue?

[Abhishek]

What methods were used to quantify mRNA and miRNA from the raw fastq files ?

[struggler]

Hi Abhishek,

For miRNA, first I predicted novel miRNA by miRDP2 and then aligning clean reads over it to get count file. 1>made database for predicted novel miRNA 2> map all filtered read over it 3> sam file to count file by samtools.4> transformed by VST 5> plot PCA For mRNA, (other person is doing) he used feature count for mRNA quantification.

Hope I could clearly explain.

[Trivas]

You might want to show examples of what the two PCA plots look like. PCA relies on the variation between your samples. Since RNA-seq has many more transcripts than your short RNA-seq, there are more instances of variation which would be more easily depicted on a PCA plot. This could potentially mean that your short RNA-seq data have little variation between conditions.

[struggler]

Hi Trivas,

Thanks for your answer. I have time series data from fresh(DT00) to dehydrated(DT03,06,09,12,15,24) to rehydrated condition(RT00,01,02,04,08,12,48) in one plant. My PCA plot shows dehydrated sample i.e DT24 comes closer to RT48 which should not be. Means both are completely opposite to each other while RNAseq PCA plot nicely distributed according to their water content. Here I am attaching PCA plot for miRNA while for mRNA can't as other person is doing. Hope I could make clear myself.