RNA Sequencing Peter Tsai Bioinformatics Institute, University of Auckland.

RNA SequencingPeter Tsai

Bioinformatics Institute, University of Auckland

Study of transcriptomes Identify known genes, exons, splicing events, ncRNA,

miRNA Novel genes or transcripts Abundances of transcripts (quantitive expression) Differential expressed transcripts between different

conditions Reconstructing transcriptome.

What is RNA-seq?

General workflowRaw data

QC

Map to reference genome

De novo transcriptome

assembly

Estimate abundance

Normalisation

Differential expression

analysis

Require downstream annotation

Use FastQC, SolexQA Trim off low quality region, keep only proper-paired reads Most QC software assume normality, but in RNA-seq data

you will probably see none-normality You might see some duplicated reads, its probably due to

highly expressed gene. Specific reference mapping tool that can map across

splice junctions between exons, i.e. Tophat Specific de novo transcriptome assembly software for

reconstruction of transcriptomes from RNA-seq data, i.e. Trinity

Quality checks and mapping

The total number of reads mapped to a gene/transcript(Count data or raw counts or digital gene expression)

Complexity of using simple counts Sequencing depth: the higher the sequencing depth, the

higher the counts Gene length: Counts are proportional to the length of the

gene times mRNA expression level Counts distribution: difference on how counts are distributed

among samples.

Expression value in RNA-seq

RPKM (Mortazavi et al, 2008)

◦ Reads Per Kilobase of exon model per Million mapped reads FPKM (Mortazavi et al, 2010)

◦ Fragments Per Kilobase of exon model per Million mapped reads

◦ Paired-end RNA-Seq experiments produce two reads per fragment, but that doesn't necessarily mean that both reads will be mappable.

Normalisation

Data exploration

Replicate 1

Repl

icat

e 2

Gene.ID/Description logFC logCPM LR PValue FDR1 2.563086301 5.07961611 28.4599795 9.57E-08 2.72E-052 4.003686266 2.330395704 28.3288251 1.02E-07 2.72E-053 2.71372512 9.704651395 25.01930526 5.68E-07 0.0001006534 -2.052703196 3.402621025 21.11492168 4.33E-06 0.0005752875 1.95117636 4.438847349 19.21195535 1.17E-05 0.0012446516 2.465833373 12.20593577 10.91756889 0.000952565 0.0844607927 1.817858683 5.308092036 10.3738524 0.001278126 0.0971375538 1.577603322 6.556675456 9.690419768 0.001852312 0.1106877669 1.20515812 4.542565518 9.670466698 0.001872537 0.110687766

10 1.233090336 10.08249873 9.289827985 0.002304298 0.12258865211 1.120581944 12.14988136 7.710102379 0.005491264 0.26557748212 1.045292369 4.913492018 7.039209923 0.00797442 0.35027053713 1.089867189 3.885246135 6.912558621 0.008559242 0.35027053714 1.353955354 2.21406615 5.976193603 0.014500264 0.55101003615 1.049933686 3.281031472 5.737563572 0.016605812 0.58895279516 -1.032999983 1.480514873 4.712476717 0.029944481 0.99565399817 -1.313778857 4.325330722 4.169234925 0.041164384 0.99874210218 0.864451602 4.338668381 3.479808135 0.062121942 0.99874210219 -0.766266641 5.2972332 3.443865378 0.063486998 0.998742102

Up-regulated

Down-regulated

Set of external RNA transcripts with known concentration. Dynamic range and lower limit of detection Fold-change response Internal control, in order to measure against defined

performance criteria

ERCC spike-in control

The dynamic range can be measured as the difference between the highest and lowest concentration.

Measure of sensitivity, and it is defined as the lowest molar amount of ERCC transcript detected in each sample

Dynamic range and lower limit of detection

Fold-change response

Depends on a number of factors◦ Biological questions

Complexity of the organism Types of analysis Types of RNA, miRNA, lncRNA.

Literature search for similar work Pilot experiment

How much library depth is needed for RNA-seq?

Have 3 or more biological replicates Analysis your data with different normalisation

methods Perform data exploration Use a standard spike-in as internal control Validation with qPCR

Summary