rna-seqbasics: from reads to differential...

26
RNA-seq basics: From reads to differential expression COMBINE RNA-seq Workshop

Upload: others

Post on 07-Jun-2020

18 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq basics:From reads to differential

expression

COMBINE RNA-seq Workshop

Page 2: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA sequencing (RNA-seq)• Use of ultra high-throughput sequencing (‘next-’ or

‘second’-generation) technologies to study gene expression

• Many applications: differential expression, transcript discovery, splice variants, allele-specific expression

• In this hands-on course, you will learn how to use statistical methods to assess differential expression in RNA-seq data using popular tools in R/Bioconductor

Page 3: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Genes and transcripts

Gene

transcript

Slide from Alicia Oshlack

Page 4: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

From transcripts to short reads

Pepke et al, Nature Methods, 2009

Page 5: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Raw data comes in fastq files• Short sequence reads• Quality scores

@HWI-ST1148:308:C694RACXX:5:1101:1768:1990 1:N:0:CGTACGNTAGGCCTTGGCAGTTTTGGAGAATCACTGCTGCCAAAGAGTCTACTTGG+#0<FFFFFFFFFFIIIIIIIIIIIIIIIIIIIIIIIIIIIIFFIIIIIII@HWI-ST1148:308:C694RACXX:5:1101:3409:1990 1:N:0:CGTACGNAGTTACCCTAGGGATAACAGCGCAATCCTATTCTAGAGTCCATATCAAC+#000BFBFFFFFFF<BFFFFBBBBBFBBFF<<FBFFIBFFFBFFFIIBFF

50 bp sequence

Page 6: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq analysis stepsRaw sequence reads

Map to genome

Summarize reads over genes

Statistical testing: Determine differentially expressed genes

Pathway analysis

Slide from Alicia Oshlack

Page 7: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Mapping reads to the genome• Where do the millions of short sequences come

from in the genome?• Sequencing transcripts, not the genome

CDS CDS CDS CDS

CDS CDS CDS CDS

Gene

transcript

Page 8: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Lots of good aligners handle splice junctions well

Exon 1 Exon 2

Page 9: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Aligned reads (bam files)HWI-ST1148:308:C694RACXX:6:2209:15171:26188 272 chr10 76314 0 50M * 0 0 CATCTGATCTTTGACAAACCTGACAAACACAAGCAATGGGGAAAGGATTC IIIIIIIIIIIIIIIIIFIFFFIIIIIIIIIIIIIIIFFFFFFFFFFBBB NH:i:10 HI:i:6 AS:i:49 nM:i:0

HWI-ST1148:308:C694RACXX:6:2306:17518:85846 272 chr10 76315 0 50M * 0 0 ATCTGATCTTTGACAAACCTGACAAACACAAGCAATGGGGAAAGGATTCC BFIIIIIIFFIFFBFFFFFFFFFFFFFIIIFFFFFFIFFFFFFFFFFBBB NH:i:10 HI:i:7 AS:i:49 nM:i:0

A row for each sequenceMillions of rows...

Page 10: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq analysis stepsRaw sequence reads

Map to genome

Summarize reads over genes

Statistical testing: Determine differentially expressed genes

Pathway analysis

Slide from Alicia Oshlack

Page 11: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Counting over exons vs counting over genes

Exon 1 Exon 2

Exon 1 = 8 readsExon 2 = 10 reads

Counting over whole gene (Exon1 + Exon2) = 15

Page 12: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Summarization turns mapped reads into a table of counts

Tag ID A1 A2 B1 B2ENSG00000124208 478 619 4830 7165ENSG00000182463 27 20 48 55ENSG00000125835 132 200 560 408ENSG00000125834 42 60 131 99ENSG00000197818 21 29 52 44ENSG00000125831 0 0 0 0ENSG00000215443 4 4 9 7ENSG00000222008 30 23 0 0ENSG00000101444 46 63 54 53ENSG00000101333 2256 2793 2702 2976

… … tens of thousands more tags …

** very high dimensional data **

Page 13: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq analysis stepsRaw sequence reads

Map to genome

Summarize reads over genes

Statistical testing: Determine differentially expressed genes

Pathway analysis

Slide from Alicia Oshlack

Page 14: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Assessing differential expression• For each gene in each sample we have a

measure of abundance – Number of reads mapping across gene

• We want to know whether there is a statistically significant difference in abundance between treatments/groups/genotypes

Page 15: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Is this gene differentially expressed?

Group 1 Group 2

Data from Shireen Lamande

Page 16: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Group 1 Group 2

outlier

Replication is really important!

Is this gene differentially expressed?

Page 17: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Quality control – check your data!

Data from Andrew Elefanty

Sorted cell populations

Page 18: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Things to think about before statistical testing

• Filtering out lowly expressed genes– Need to make decisions about cut-offs – Can be an iterative process

Want to avoid calling this gene DE due to one sample

Page 19: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

• Normalisation– Library size (sequencing depth)– Composition bias (TMM)– Batch effects (RUVSeq)

Things to think about before statistical testing

.

(a)

log2(Kidney1 NK1) − log2(Kidney2 NK2)

Den

sity

-6 -4 -2 0 2 4 6

0.0

0.4

0.8

log2(Liver NL) - log2(Kidney NK)

Den

sity

-6 -4 -2 0 2 4 6

0.0

0.2

0.4(b)

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

●●

●●●

● ●●

●●

●●

●●

● ●●

● ●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

● ●

●●

● ●

●●

● ●

●●

●●

●●

● ●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

● ●

●●

●●●

●●

● ●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●● ●

● ●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

● ●●

●●

●● ●

●●

●●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●● ●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●●

● ●

●●

●●

●●

●● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●● ●●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

● ●

●●● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

● ●

●●

● ●

●● ●

● ●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

● ●

● ●

● ●

●●

●● ●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

● ● ●

● ●

● ●●

● ● ●

●●●

●● ●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

● ●●

●● ●

●●

●●

● ●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●●●

● ●

●●

●●

●● ●●

● ●

●●

●●

●●●

●●●

● ●

●●

● ●

●●

●●

● ●●

●●●

●●

●●

●●

●● ●●

● ●●

●●

●●

●●

●●

●●

●● ●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●●

●● ●●● ● ●

●●●

●●

●● ●

● ●

● ●

●●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●● ●

● ●

●●

●●

●●

● ●

● ●

●●

●●

● ●

● ●

●●

●● ● ●

●●

●●

●●

●●

●●

●●

●●

● ●● ●

● ●●

● ●●

●● ●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

● ●

● ●

●●

●●

● ●

●●

● ●

●●

●●

● ●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●● ●

●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●●

●●

● ●

●●

● ●

● ●

●●

●●

● ●

●●

● ●

●●

●● ●

●●●●

●●●

●●

●●

●●

● ●

●●

●●

● ●●

● ●

●● ●●

●●

●●

●●

● ●

●●

●●

● ●

● ● ●

●●

●●

● ●●

●●

●●● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●● ●

●●

●●●

●●●●

●●

● ●

●●

●●● ●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●●

● ●

●● ●

● ●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●● ●

●●

●●

● ●

●●●

● ●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

● ●●

●●

●●

●●

●●

●●●

●● ●

● ●

● ●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●●

● ●

●●

● ●

●●

●●

● ●

●●

● ●

●●

●●

●● ●●

●●

●●

●●

●●

●●

● ● ●●

●●

●●

●●

● ● ●●

●●

● ●

●●

●●

● ●

●●

● ●

●●

●●

● ●

●●

●●

● ●●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

● ●

●●●●

●●

●● ●

●●

●●

●● ●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

●●

● ●●

●●

● ●●

● ●●

●●

●●

●●

● ●

●●

●●

●●

●● ●●

●●

●●

●● ●

●● ●

● ●

●●

●●

●●

● ●

●● ●

●●

● ●●

● ●

●●

●●

● ● ●

●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

● ● ●

● ●

●●

●●

●●

●●

●●

● ●

● ●●

●●

● ●

●●● ●● ●

●●

●●

●●●

●●●

●●

●●

●●

●●●

●●

● ●

● ●

●●

●●

●●

●●●

● ● ●

●●

●●

●●

●● ●

●●●

● ●

●●●

●●

● ●●

●●

●●

● ●●

●●●●●

●●●

●●

●●

●●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

● ●●

● ●

● ●

●●●

● ●●

● ●

●●

● ●

●●

●●

●●

●●

● ●

● ●

● ●

●●

●●

●●

● ●

● ●

●●

●●

●● ●

●●

●●

●●

● ●● ●

● ●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

● ● ●

●●

●●

●● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

● ●●

● ●

● ●●

●●

●●

● ●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●●

●● ●●

●●

●●

●●

●●

●●

●●

● ●●

●●● ●

● ●

●●

●●

●●

● ●

●●

●●● ●

●●

● ●

●●

●●

●●

●● ●●●

●●

●●

●●

● ●

●●

●● ●

● ●●

●●

●●

●● ●

●●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●

●● ●

● ●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

● ●●

●●

●● ●

●●

●●

●●

●●

● ●

●● ●●

●●

● ●

● ●

● ●

●●●

●●●

● ●

●●

● ●

●●

● ●

●●

●●

●● ●

● ●

●●

●●● ● ●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●● ● ●●●

●●

● ●●

●●●

● ● ●●

●●

●●

●●●●

●●

●● ●

●●

●●

● ●●

●●

●●●●

● ●●●

●●

●●

●●

●● ●

● ●●

●●

●●

●●

●●

●●●

● ● ●

●●

●●

● ●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

● ●

●●●

●●

● ●

●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●● ●

●●

●●

●●

●●

● ●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

● ●

● ●

● ●

● ●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●● ●

● ●

●●

●●

●●

●●

●●

●● ●●● ●

● ●

● ●

●●●

● ●

●●

●●

●●

● ●●●

●●●●

●● ●

● ● ●

●● ●

●●

●● ●

●●

●●

●●

●●

●●●●

● ●

● ●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●● ●

●●●

● ● ●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●● ●

●●●

●●

●●

●●

● ●

●●

●●

●●

●● ●

●● ●

●●

●●

● ●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●●

●●

● ●

● ●

●●

●●

●●●

●●

● ●

●●●

●●

●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

● ●

● ●

●●

● ●

● ●

●● ●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

● ●

● ●

●●●●

●●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●●

● ●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●● ●

●●

●●

● ●

● ●

●●

●●

● ●

● ●

●●

●●

●●

●●

● ●

● ●

●●●

●●

●●

●●

●●

●● ●

●●

● ●

●●

●●

● ●

●●

●●

● ●

● ●

●●●

●●

●●

●●

●●

● ●●

● ●

●●

●●●

●●

● ●

● ●

●●

● ●

●●

●●

●●

●● ●

●●

●●

● ●

●●

●●

●●

●●● ●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●●

●●

●●

● ●

● ●

●●

●●

● ●

●●

●●●●

●●

●●

●●

●● ●

● ●

●●

●● ●

●●

● ●

●●

●●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ● ●

●● ●

●●

●●

● ●● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

● ●●

●●

●●

●●

●●

● ●

●●●

●●

● ●

●●

●●

●●

● ●

● ●

●●

●●

●●

● ●

●●

● ●

●●●

●● ●

● ●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●● ●

●●●● ●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●● ●

● ●

● ●

●● ●●

●●

●●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

● ●● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●●

●●

●●

●●

● ●●

● ●

● ●●●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

● ●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●● ●

●●

●●

● ●●●

●●

● ●

●●

●●

● ●

● ●●

●●

●●

●●

●●

●●

●●●

●● ●● ●

● ●●

●●

●● ●● ●●

● ●

●●

● ●

●●

●●

●●

●●

●●

● ●

● ●

●●●

●● ●

●●

● ●

●●

● ●

● ●●

●●

●●

●●

●●

●● ●

●●

●●

● ●

●●

●●

● ●

●●

●● ●

●●

● ● ●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

● ●●

●●

● ●

● ●●

●●

●●

● ●

●●

●●

●●

● ●●

●●

● ●

● ●

●●

●●

●●●

●●

●● ●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●● ●

● ●

● ●

●●

●●

●●

●●

●●●

●●●

● ●●●● ●●

●●

●●

●●

●●● ● ●

● ●

●●●

●●

● ●

●● ●

●●

●●

●●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●●

●●

●●

●●

● ● ●

●●

●●

● ●

●● ●

● ●

● ●

●●

● ●

●●

●●

● ●●

●●

●●

●●

●●

●●

● ●

● ●

●●●

●●●●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

● ●

● ●

●● ●●

●●

●●

● ●

● ●

●●

●●

● ●

●●●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●

●●

●●

●●

● ●

● ●

●●

● ●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●●

●●

●●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●●●

●●●

●●

●● ●

● ●

● ● ●

●●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

● ●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●●

●●

●●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●●●

●●●

● ●●

● ●

●●

●● ●●

●●

● ●●

●● ●

●●●

●●

●●●

●●●

● ●

● ●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

● ●

● ●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

● ●

●●

● ●

●●

●● ●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●● ●

●●

●●

●●

● ●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●● ●

●●●●●

●●

●●

●●

●●●

●●

●●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

● ●●

●●

● ●●

●●

●●

●●

● ●

●●

● ●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

● ●●

●●

● ●

● ●

●●

●●

●●

●●

●●

● ●

●● ●

●●

● ●

●●

●●

● ●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

● ●

● ●

●●

● ●

● ●

● ●

●● ● ●

●●

●●●

●●

● ●●

●●

●●

●●

● ●●●

●●

● ●●

● ●

●●

● ●

● ●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●● ● ●

● ●●

● ●

●●

●●

●●

● ●

● ●

● ●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●● ●

● ●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●●

●●

●●

●●●

●● ●

●●

●●

●●

●●

●●

●●

● ●● ●

●●

●●

●●

● ●●

●●

● ●

●●

●●

●●

● ●

●●

●●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●●● ●●

●●

● ●

● ●● ● ●●

● ●

●●

●●

●●

●●●●

●●

●●

●● ●

● ●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

● ●

●●

●●

●● ●●

● ●

●● ●

●●●

● ●

● ●

●●

●●

●●

●● ●

● ● ●

●●

● ●

●●

●●

●●

● ● ●●

●●

●●●

●●

● ●

●●

●●

● ● ●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●●

●●

●● ●

●●

●●

●●

●●

● ●● ●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

● ●

●●●

●●

● ●● ● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●●

●●●●

●● ●

●● ●

●●

●●

●●

●● ●

●●

●●

●●

● ●

●●

●●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●

●●

● ●

●●

●●

●●

● ●

●●

●● ●

●●

● ●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●

●●

●●

●● ●

● ●

●●

●●

●●●

●●

●●●

●●

● ●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●● ●

●●

●●

● ● ●

●●

-20 -15 -10

-50

5

A = log2( Liver NL Kidney NK)

M=

log 2

(Liv

erN

L)-

log 2

(Kid

ney

NK)

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●● ●

●●

●●

● ●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●●

●●

● ●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

Housekeeping genesUnique to a sample

(c)

Page 20: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Statistical testing for DE

• For each gene, is the mean expression levelunder one condition significantly different from the mean expression level under a different condition?

Tag ID A1 A2 B1 B2ENSG00000124208 478 619 4830 7165

ENSG00000182463 27 20 48 55

ENSG00000125835 132 200 560 408ENSG00000125834 42 60 131 99

… … tens of thousands more tags …

20

Page 21: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Many different statistical methods• Model the counts directly– Negative binomial modelling is best because it

captures biological as well as technical variability– Most popular packages in R• edgeR• DESeq/DESeq2• Lots of others exist (baySeq, NBPSeq, …)

• Transform the counts and used normal based methods– voom in the limma package

Page 22: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Statistical testing gives each gene a p-value for evidence of DE

Tag ID A1 A2 B1 B2ENSG00000124208 478 619 4830 7165ENSG00000182463 27 20 48 55ENSG00000125835 132 200 560 408ENSG00000125834 42 60 131 99ENSG00000197818 21 29 52 44ENSG00000125831 0 0 0 0ENSG00000215443 4 4 9 7ENSG00000222008 30 23 0 0ENSG00000101444 46 63 54 53ENSG00000101333 2256 2793 2702 2976

… … tens of thousands more tags …

Tag ID P-valueENSG00000124208 0.0002ENSG00000182463 0.12ENSG00000125835 0.034ENSG00000125834 0.08ENSG00000197818 0.64ENSG00000125831 1ENSG00000215443 1ENSG00000222008 0.06ENSG00000101444 0.73ENSG00000101333 0.22

Page 23: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq analysis stepsRaw sequence reads

Map to genome

Summarize reads over genes

Statistical testing: Determine differentially expressed genes

Pathway analysis

Slide from Alicia Oshlack

Learn something!

Page 24: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

Summary• Lots of choices in analysis methodology• Quality control is essential! Sometimes

detective work is necessary.• Each step of the analysis requires decisions

that impact down-stream analysis• Life gets harder when there’s no genome or

poor quality genomes

Page 25: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

RNA-seq analysis in R / Bioconductor

Page 26: RNA-seqbasics: From reads to differential expressioncombine-australia.github.io/RNAseq-R/slides/RNASeq_basics.pdf · RNA sequencing (RNA-seq) •Use of ultra high-throughput sequencing

AcknowledgementsSlides:• Alicia Oshlack• Belinda Phipson• Anthony Hawkins• Gordon Smyth• Davis McCarthy

Data:• Andrew Elefanty and Elizabeth Ng• Shireen Lamande