Relationship between trimming read data set and expression value in RNA-seq

[bongbimit]

Hi everyone,

I have a question about RNA-seq trimming of raw read data. After using Illumina Miseq teachnology to sequence my mRNA sample, my raw read data set was trimming to filter low quality (Q<30) and I obtained a clean read data set (80% of total clean read). Then I do mapping and normalization expression value by Tophat and Cuffdiff.

So I have a question: If I have a better sequencing result (I mean i do not trimming 20% low quality raw read), do I have a different expression value on some gene?! Because as I know, Cuffdiff using FPKM or RPKM as a unit expression value and these units were affected by total read count. Therefore do I have a truly expression with the clean read data set.

Thank all.

[Brian Bushnell]

Trimming/filtering to Q30 is ridiculously high. I doubt that's ever beneficial, and it will certainly cause bias in quantitative analyses.