Population scale transcriptomics for

precision oncology

​

​

Jeff Leek

VP & CDO

Professor Biostatistics Program, PHS

J Orin Edson Foundation Chair

@jtleek

​

1

2

3

Fred Hutchinson Cancer Center

www.github.com/jtleek/talks

Fred Hutchinson Cancer Center

4

Precision oncology:

A tour of data plumbing challenges

​

​

Jeff Leek

VP & CDO

Professor Biostatistics Program, PHS

J Orin Edson Foundation Chair

@jtleek

​

5

6

Fred Hutchinson Cancer Center

What gene expression patterns are prognostic of colorectal cancer metastasis ?

7

Fred Hutchinson Cancer Center

https://pubmed.ncbi.nlm.nih.gov/25049118/

Q: What genes are prognostic of colorectal cancer metastasis?

​

Data

Processing

Computing

Metadata

Analysis

A: This gene signature is/isn’t a potential prognostic biomarker.

​

9

Fred Hutchinson Cancer Center

Find a researcher with access to patient samples

What genes are prognostic for metastasis?

Find a researcher with access to patient samples

What genes are prognostic for metastasis?

Collect patient samples and information

3~6 months

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

3~6 months

1-2 wks

2-4 wks

What genes are prognostic for metastasis?

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

Process sequencing data

33-6 months

1-3 months

1 month - 1+ years

3~6 months

1-2 wks

2-4 wks

Analyze data and answer biological question

Data cleaning

What genes are prognostic for metastasis?

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

Process sequencing data

33-6 months

1-3 months

1 month - 1+ years

3~6 months

1-2 wks

2-4 wks

Total: 2+ years

Analyze data and answer biological question

Data cleaning

What genes are prognostic for metastasis?

Data

Processing

Computing

Metadata

Analysis

16

Fred Hutchinson Cancer Center

We are at the intersection of two revolutions

17

Fred Hutchinson Cancer Center

18

Fred Hutchinson Cancer Center

SAMPLE SIZE

N =

19

Fred Hutchinson Cancer Center

N =

($ YOU HAVE)

($ PER SAMPLE)

20

Fred Hutchinson Cancer Center

Langmead & Nellore, Nat Rev. Genet. 2018

21

Fred Hutchinson Cancer Center

http://www.washingtonpost.com/sf/national/2015/06/27/watsons-next-feat-taking-on-cancer/

​

Data sharing is improving over time

​

...but data aren’t always easy to use

​

Data sharing is improving over time

​

Data

Processing

Computing

Metadata

Analysis

26

Fred Hutchinson Cancer Center

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

Process sequencing data

33-6 months

1-3 months

1 month - 1+ years

~6 months

1-2 wks

2-4 wks

Total: 1.5+ years

Analyze data and answer biological question

Data cleaning

What genes are prognostic for metastasis?

AUCAGUCGAUCACCGAU

transcription

RNA

translation

protein

ACTGACCTAGATCAGTCGATCGATCGTATACGATTACAAAATCATCGGCAT

DNA

M

M

M

slide adapted from alyssa frazee

AUCAGUCGAUCACCGAU

transcription

RNA

translation

protein

ACTGACCTAGATCAGTCGATCGATCGTATACGATTACAAAATCATCGGCAT

DNA

M

M

M

RNA

slide adapted from alyssa frazee

Genome

Transcripts

Reads

@22:16362385-16362561W:ENST00000440999:2:177:-40:244:S/2

CCAGCCCACCTGAGGCTTCTTTTTCCTTCCCAAGCCACATCACCATCCTGGTGGAACTCTCCTGTGAGGACAGCCA

+

GGFF<BB=>GBGIIIIIIIIIIIIIIEGEHGHHIIIIIIIIHFHBB2/:=??EGGGEGFHHIHHEDBD?@@DDHHD

@22:16362385-16362561W:ENST00000440999:3:177:-56:294:S/2

GCGTGAGCCACAGGGCCCAGCCCACCTGAGGCTTCTTTTTCCTTCCCAAGCCACATCACCATCCTGGTGGAACTCT

+

@=ABBBBIIIIIIIIHHGGGGIIDBDIIIIIIGIIIIHIIIIHFDD@BBDBGGFIDEE8DCC/29>BGFCGHHHGF

@22:16362385-16362561W:ENST00000440999:4:177:137:254:S/1

TCACCATCCTGGTGGAACTCTCCTGTGAGGACAGCCAAGGCCTGAACTACCTGCaGTGGGGAGCACCTCAGGGTTT

+

DDGBBCGGGIGGGBDDDHIIGGDGD77=BDIIIIIIIIFHHHHIIIHEFFHGGDD8A>DEGHHIFDDHH8@BEDDI

@22:16362385-16362561W:ENST00000440999:5:177:68:251:S/2

AGGGTTTGCCCAGGCAACCAGCCAGCCCTGGTCCAAGGCATCCTGGAGCGAGTTGTGGATGGCAAAAAGACNCGCC

+

HIGHIHFHEGE4111:.;8@?@HDIIIIIIIEGGIHHHIIGA?=:FIIIDD8.02506A8=AC#############

@22:16362385-16362561W:ENST00000440999:6:177:348:453:S/1

AAGGCCTGAACTACCTGCGGTGGGGAGCACCTCAGGGTTTGCCCAGGCAACCAGCCAGCCCTGGTCCAAGGCATCC

+

B9?@8=42:E@GDEDIIIIIGGHIIIFBEEAGIIDIIDHHGGHIIEGEIIIIIHIHFHFFEEFGGGGGB88>:DGH

@22:51205934-51222090C:ENST00000464740:132:612:223:359:S/2

GGAAGTATGATGCTGATGACAACGTGAAGATCATCTGCCTGGGAGACAGCGCAGTGGGCAAATCCAAACTCATGGA

+

IIEHHHHHIIIIIIIHGGDGHHEDDG8=;?==19;<<>D@@GGGIIHIIHGGDDHGBA=ABEG@@DFCCAA<:=>8

@22:51205934-51222090C:ENST00000464740:125:612:-1:185:S/1

TGGAGTGCGCTGCGGCGCGAGCTGGGCCGGCGGGCGTGGTTCGAGAGCGCGCAGAGTCCAGACTGGCGGCAGGGCC

+

HHIIIHIDGG@;=@GIIIIIDDGBBBEDB@8>5554,/':9B@@C?==@1:2@?=GG=;<HHHHGIIHHEC-;;3?

​

​

3 gb

Carefully!!

coverage vector

2

6

0

11

6

Genome

(DNA)

junction

counts

3

3

Genome

(DNA)

~71.7 mb

Data

Processing

Computing

Metadata

Analysis

38

Fred Hutchinson Cancer Center

SRA

Human

RNA-seq

Illumina

≈22,000 samples

≈50,000 samples

≈60,000 samples

≈247,000 samples

​

​

http://rail.bio/

Slide courtesy Ben Langmead

http://blogs.citrix.com/2012/10/17/announcing-general-availability-of-sharefile-with-storagezones/

slide adapted from andrew jaffe

Obstacle: our research moves (spot) markets

Spike in market price due to preprocessing job flows

�expression data for ~70,000 human samples

GTEx

N=9,962

TCGA

N=11,284

SRA

N=49,848

samples

expression estimates

gene

exon

junctions

ERs

< 1,000 junctions

> 20,000 junctions

=

A global view of transcript variability

Nellore et al. Genome Biology 2016

47

Fred Hutchinson Cancer Center

We discovered and redefined hundreds of human genes using these data!

48

Fred Hutchinson Cancer Center

49

Fred Hutchinson Cancer Center

Data

Processing

Computing

Metadata

Analysis

50

Fred Hutchinson Cancer Center

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

Process sequencing data

33-6 months

1-3 months

1 month - 1+ years

~6 months

1-2 wks

2-4 wks

Total: 1.5+ years

Analyze data and answer biological question

Data cleaning

What genes are prognostic for metastasis?

What % expressed?

New genes?

Important outliers?

Prognostic signatures

Sequence data

Process/quantify

ACTACTTT

Metadata

Clean/predict

�expression data for ~70,000 human samples

GTEx

N=9,962

TCGA

N=11,284

SRA

N=49,848

samples

expression estimates

gene

exon

junctions

ERs

Answer meaningful questions about human biology and expression

�expression data for ~70,000 human samples

samples

phenotypes

?

GTEx

N=9,962

TCGA

N=11,284

SRA

N=49,848

samples

expression estimates

gene

exon

junctions

ERs

Answer meaningful questions about human biology and expression

SRA phenotype information is far from complete

z

z

z

Even when information is provided, it’s not always clear…

Sex across the SRA:

Even when information is provided, it’s not always clear…

Sex across the SRA:

“1 Male, 2 Female”, “2 Male, 1 Female”, “3 Female”, “DK”, “male and female” “Male (note: ….)”, “missing”, “mixed”, “mixture”, “N/A”, “Not available”, “not applicable”, “not collected”, “not determined”, “pooled male and female”, “U”, “unknown”, “Unknown”

Goal :��to accurately predict critical phenotype information for all samples in recount2

​

gene, exon, exon-exon junction and expressed region RNA-Seq data

SRA

Sequence Read Archive

N=49,848

​

TCGA

The Cancer Genome Atlas

N=11,284

GTEx

Genotype Tissue Expression Project

N=9,662

Missingness limited in GTEx phenotype data

GTEx

Sex across GTEx:

Goal :��to accurately predict critical phenotype information for all samples in recount2

Ellis et al. Nuc. Acids Res. 2018

​

gene, exon, exon-exon junction and expressed region RNA-Seq data

SRA

Sequence Read Archive

N=49,848

​

divide samples

build and optimize phenotype predictor

predict phenotypes across SRA samples

test accuracy of predictor

TCGA

The Cancer Genome Atlas

N=11,284

GTEx

Genotype Tissue Expression Project

N=9,662

predict phenotypes across samples in TCGA

Training Data

Validation

Data

Test

Data

Validation

Data

99.9%

Sex prediction is accurate across data sets

99.8%

99.0%

86.3%

To assess misreporting of sex in the SRA, we can use Y-chromosome expression

XX

XY

reported female

reported male

​

predicted female

predicted male

64

Fred Hutchinson Cancer Center

Expression from the Y chromosome suggests misreporting of sex in the SRA

65

Fred Hutchinson Cancer Center

Ellis et al. Nuc. Acids Res. 2018

�expression data for ~70,000 human samples

samples

phenotypes

GTEx

N=9,962

TCGA

N=11,284

SRA

N=49,848

samples

expression estimates

gene

exon

junctions

ERs

Answer meaningful questions about human biology and expression

What % expressed?

New genes?

Important outliers?

Sequence data

Process/quantify

Metadata

Clean/predict

ACTACTTT

Primary vs.

Metastatic

What % expressed?

New genes?

Important outliers?

Sequence data

Process/quantify

Metadata

Clean/predict

ACTACTTT

Primary vs.

Metastatic

​

​

​

​

ACTATTTT

Alt Sequence

Data

Processing

Computing

Metadata

Analysis

73

Fred Hutchinson Cancer Center

Find a researcher with access to patient samples

Collect patient samples and information

Extract DNA/RNA from samples

Sequence samples

Process sequencing data

3-6 months

1-3 months

1 month - 1+ years

~6 months

1-2 wks

2-4 wks

Total: 1mo - 1+years

Analyze data and answer biological question

Data cleaning

What genes are prognostic for metastasis?

What gene expression patterns are prognostic of colorectal cancer metastasis ?

75

Fred Hutchinson Cancer Center

https://pubmed.ncbi.nlm.nih.gov/25049118/

Kim et al. analysis looked to identify genes that contribute to metastasis in colon cancer.

N=18

3. Liver Metastasis (MC)

1. Healthy Colon (NC)

2. Primary Cancer (PC)

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

Are the same genes found when sex is included in the analysis?

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

Concordance at the Top (CAT) Plots

How similar are the results from Analysis A and Analysis B ?

Analysis A

Analysis B

80

Fred Hutchinson Cancer Center

Concordance at the Top (CAT) Plots

How similar are the results from Analysis A and Analysis B ?

​

Analysis A

Analysis B

10

10

81

Fred Hutchinson Cancer Center

Concordance at the Top (CAT) Plots

How similar are the results from Analysis A and Analysis B ?

​

Analysis A

Analysis B

10

10

82

Fred Hutchinson Cancer Center

Concordance at the Top (CAT) Plots

How similar are the results from Analysis A and Analysis B ?

​

Analysis A

Analysis B

The top results from A and B are the same

Some differences at less significant genes

83

Fred Hutchinson Cancer Center

Concordance at the Top (CAT) Plots

How similar are the results from Analysis A and Analysis B ?

Analysis A

Analysis B

The results of the orange condition are less similar between Analysis A and B than the green condition

84

Fred Hutchinson Cancer Center

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

NC: non-cancerous

PC: primary cancer

MC: metastatic cancer

Predictions can be used to:��(1) Identify studies of interest��(2) appropriately analyze data

Loss of concordance suggests that differential expression is detecting tissue differences, not cancer-related changes.

We have expression data from both healthy liver and colon samples (GTEx)...

90

Fred Hutchinson Cancer Center

So…what if we compared the MC:PC results with differential expression between colon and liver?

Hypothesis: MC:PC results should be most similar to GTEx colon vs. liver

91

Fred Hutchinson Cancer Center

Comparison of results with GTEx colon vs. liver suggests differential expression results detecting tissue differences

92

Fred Hutchinson Cancer Center

Data

Processing

Computing

Metadata

Analysis

93

Fred Hutchinson Cancer Center

Data

Processing

Computing

Metadata

Analysis

Software

94

Fred Hutchinson Cancer Center

Study explorer

https://jhubiostatistics.shinyapps.io/recount3-study-explorer/

95

Fred Hutchinson Cancer Center

Bioconductor packages

recount3

snaptron

dasper

96

Fred Hutchinson Cancer Center

Data

Processing

Computing

Metadata

Analysis

Software

Training

97

Fred Hutchinson Cancer Center

General purpose training

https://www.coursera.org/specializations/genomic-data-science#instructors

98

Fred Hutchinson Cancer Center

recount workshops

http://research.libd.org/recountWorkshop/

99

Fred Hutchinson Cancer Center

Thanks to many, many people!

Ben Langmead

Kasper Hansen

Abhi Nellore

Chris Wilkes

Leo Collado Torres

Shannon Ellis

Afrooz Razi

Chris Lo

​

​

Alyssa Frazee

Feng Young Chen

Rone Charles

Kai Kammers

Margaret Taub

Jack Fu

• Many others!

​

Siruo Wang

Andrew Jaffe

Shije Zheng

Lance Joseph

David Zhang

Eddie Luidy Imada�Jonathan Ling

Brad Solomon

100

Fred Hutchinson Cancer Center

101

Fred Hutchinson Cancer Center

Thank you