8 All about RNAseq

8.1 General

  • Functional annotation of the mouse (FANTOM)
  • Encyclopedia of DNA elements (ENCODE)
  • Single loci can lead to to six different mRNA isoforms, on average, instead of single transcript

8.2 Technology

What is it for?

  • what is expressed (exon locatios /strand) ?
  • how much is expressed (hits per transcripts) ?

old way to do DE - microarray

  • do need as much starting material for RNAseq

Noise:

  • biological
  • poisson noise from samplinig (have level of uncertanty at low level, low counts)

  • technical

  • Effect size (do the genes go up by alot or not - high signal)

  • amount of variation

get better depth using poly-A enriched, because rRNA depletion will leave lots of other RNA around therefore diluting mRNA by some percent

mean fragment size ~ 300 bp

measure of uncertanty of the base call

8.3 Differential expression (DE)

  • library size is total number of mapped reads. Library size might vary for individual samples.

  • Total number of reads per gene proportional to:

    • gene expression level (to get cpm (count per million) divide each read count by library size multiplied by million). This is enable you to compare different library sized samples
    • transcript length
    • sequencing depth of the library # Ideas
  • have a url with multiqc report ready for browsing

  • explain reference file (FASTA) and annotation files (GTF)