machine reading for cancer panomics - akbc · 2018. 11. 4. · genomics 12 discovery …...

Machine Reading for

Cancer Panomics

Hoifung Poon

1

Overview

2

… ATTCGGATATTTAAGGC …

… ATTCGGGTATTTAAGCC …

……

……

Disease Genes

Drug Targets

……KB

Cancer Systems Modeling

High-Throughput Data

3



……

……

Disease Genes

Drug Targets

…KB

Extract Pathways

from PubMed

Overview

High-Throughput Data

Grounded

Semantic Parsing

Precision Medicine

5

Before Treatment 15 Weeks

Vemurafenib on BRAF-V600 Melanoma

Vemurafenib on BRAF-V600 Melanoma

6

Before Treatment 15 Weeks 23 Weeks

Traditional Biology

8

Targeted Experiments Discovery

One

hypothesis

Genomics

9

High-Throughput ExperimentsDiscovery







Many

hypotheses

?


… ATTCGGGTATTTAAGCC … Healthy

Disease(e.g., Alzheimer, Cancer)

Genome-Wide Association Studies (GWAS)

2000

2010

“Genetic diagnosis of diseases would be

accomplished in 10 years and that

treatments would start to roll out perhaps

five years after that.”

“A Decade Later, Genetic Maps Yield Few New Cures”

New York Times, June 2010.

10

Key Challenges

Human genome: 3 billion base pairs

Potential variations: > 10 million variants

Combination: > 101000000 (1 million zeros)

Machine learning problem

Atomic features: > 10 million

Feature combination: Too many to enumerate

11

Genomics

12

Discovery







How to Scale Discovery?

High-Throughput Experiments

Cancer

Hundreds of mutations

Most are “passenger”, not driver

Can we identify likely drivers?

13


… ATTCGGGTATTTAAGCC … Normal cells

Tumor cells

Panomics

14


Genome Transcriptome Epigenome

……

Pathway Knowledge

Genes work synergistically in pathways

15

Why Hard to Identify Drivers?

Complex diseases Perturb multiple pathways

16Hanahan & Weinberg [Cell 2011]

Why Cancer Comes Back?

Subtypes with alternative pathway profile

Compensatory pathways can be activated

17

EphA2 EphB2

Ovarian Cancer

Why Cancer Comes Back?

Subtypes with alternative pathway profile

Compensatory pathways can be activated

18

EphA2 EphB2

Ovarian Cancer

X

Cancer Systems Modeling

19

Gene A DNA mRNA Protein Protein Active

Transcription Translation Activation


Functional activity

Mutation effect

Drug Target

……

20


Gene B DNA mRNA Protein Protein Active

Gene C DNA mRNA Protein Protein Active

Transcription Factor

Protein Kinase

Knowledge Model

21





Protein Kinase

?Knowledge Model

22





Protein Kinase

?Knowledge Model

23





Protein Kinase

!Knowledge Model

Approach: Graph HMM

24



Protein Kinase



Extract Pathways from PubMed

25



……

……

Disease Genes

Drug Targets

……KBHigh-Throughput Data

PubMed

24 millions abstracts

Two new abstracts every minute

Adds over one million every year

26

…

VDR+ binds to

SMAD3 to form

…

…

JUN expression

is induced by

SMAD3/4

…

PMID: 123

PMID: 456

……

27

Machine Reading

28

Involvement of p70(S6)-kinase activation in IL-10

up-regulation in human monocytes by gp41 envelope

protein of human immunodeficiency virus type 1 ...

Machine Reading

29




IL-10human

monocytegp41 p70(S6)-kinase

Machine Reading

PROTEINPROTEINPROTEINCELL

30




Involvement

up-regulation

IL-10human

monocyte

SiteTheme Cause

gp41 p70(S6)-kinase

activation

Theme Cause

Theme

Machine Reading

REGULATION

REGULATION REGULATION


31




Involvement

up-regulation

IL-10human

monocyte

SiteTheme Cause

gp41 p70(S6)-kinase

activation

Theme Cause

Theme

Machine Reading

REGULATION

REGULATION REGULATION


Semantic Parsing

Long Tail of Variations

32

TP53 inhibits BCL2.

Tumor suppressor P53 down-regulates the activity of BCL-2 proteins.

BCL2 transcription is suppressed by P53 expression.

The inhibition of B-cell CLL/Lymphoma 2 expression by TP53 …

……

Bottleneck: Annotated Examples

GENIA (BioNLP Shared Task 2009-2013)

1999 abstracts

MeSH: human, blood cell, transcription factor

Challenge for “supervised” machine learning

Can we breach this bottleneck?

33

Free Lunch #1:

Distributional Similarity

Similar context Probably similar meaning

Annotation as latent variables

Textual expression Recursive clusters

Unsupervised semantic parsing

34

Poon & Domingos, “Unsupervised Semantic Parsing”.

EMNLP 2009. Best Paper Award.

Recursive Clustering

35

TP53 inhibits BCL2.




……


36

TP53 inhibits BCL2.




……


37

TP53 inhibits BCL2.




……


38

TP53 inhibits BCL2.




……

BCL2, BCL-2 proteins,

B-cell CLL/Lymphoma 2

……

TP53,Tumor

suppressor P53

……

inhibits, down-regulates,

suppresses, inhibition, …

Theme Cause

Free Lunch #2:

Existing KBs

Many KBs available

Gene/Protein: GeneBank, UniProt, …

Pathways: NCI, Reactome, KEGG, BioCarta, …

Annotation as latent variables

Textual expression Table, column, join, …

Grounded semantic parsing

39

Entity Extraction

40

ID Symbol Alias

990 BCL2 B-cell CLL/Lymphoma 2, …

11998 TP53 Tumor suppressor P53, …

… … …

HGNC

Entity Extraction

41

ID Symbol Alias

990 BCL2 B-cell CLL/Lymphoma 2, …

11998 TP53 Tumor suppressor P53, …

… … …

HGNC

TP53 inhibits BCL2.




……

Relation Extraction

42

Regulation Theme Cause

Positive A2M FOXO1

Positive ABCB1 TP53

Negative BCL2 TP53

… … …

NCI-PID

Pathway KB

TP53 inhibits BCL2.




……

Relation Extraction

43

Regulation Theme Cause

Positive A2M FOXO1

Positive ABCB1 TP53

Negative BCL2 TP53

… … …

NCI-PID

Pathway KB

TP53 inhibits BCL2.




……

Grounded Learning

Question Answering w.r.t. KB

44

Poon, “Grounded Unsupervised Semantic Parsing”. ACL 2013.

System Accuracy

ZC07 84.6

FUBL 82.8

GUSP 83.5

Supervised

Unsupervised

Pathway Extraction

Generalize distant supervision:

Nested events in KB likely occur in

semantic parse of some sentence

Prior: Favor semantic parse grounded in KB

Outperformed the majority of participants in

original GENIA Event Shared Task

45

Parikh, Poon, Toutanova. In Progress.

http://literome.azurewebsites.net

46

Literome

Poon et al., “Literome: PubMed-Scale Genomic Knowledge

Base in the Cloud”, Bioinformatics 2014.

PubMed-Scale Extraction

Preliminary pass:

2 million instances

13,000 genes, 870,000 unique regulations

Applications:

UCSC Genome Browser, MSR Interactions Track

Expression profile modeling

Validate de novo pathway prediction

Etc.

47

Poon, Toutanova, Quirk, “Distant Supervision for Cancer

Pathway Extraction from Text”. PSB 2015. To appear.

Machine Science

48

Evans & Rzhetsky, “Machine Science”.

Science, Vol. 329, 2010.

Machine Science

49

Big Data

Machine Science

50

Big Data Rich Knowledge

KB

Machine Science

51

Deep Model


KB

Machine Science

52

Deep Model


Hypotheses

KB

Machine Science

53

Deep Model


Hypotheses

Experiments

KB

Machine Science

54

Deep Model


Hypotheses

Experiments

KB

Roadmap

Extract richer knowledge:

Cell type, experimental condition, …

Hedging, negation, …

Formulate coherent models:

Supporting evidence, contradiction, …

Intellectual gaps, hypotheses, …

Integrate w. data & experiments:

Cancer panomics Driver genes / pathways

Single-drug response Drug combo prioritization

55

Big Mechanism

42-million program

Reading, Assembly, Explanation

Domain: Cancer signaling pathways

We are in

PI: Andrey Rzhetsky

Co-PI w. James Evans, Ross King

56

57

Berkeley

AMP Lab

OHSU

Microsoft

Research

We Have Digitized Life

58

Next: Digitize Medicine

59

Knock down genes A, B, C → Cure

Summary

Precision medicine is the future

Cancer systems modeling

Graphical model: Pathways + Panomics data

Extract pathways from PubMed

Machine reading by grounded semantic parsing

Literome: KB for genomic medicine

60

Acknowledgments

61

U. Chicago: Andrey Rzhetsky, Kevin White

OHSU: Brian Drucker, Jeff Tyner

Berkeley AMP Lab: David Patterson

U. Wisconsin: Anthony Gitter

Microsoft Research: Chris Quirk, Kristina

Toutanova, David Heckerman, Ankur Parikh,

Lucy Vanderwende, Bill Bolosky, Ravi Pandya

Summary

62



……

……

Disease Genes

Drug Targets

……KBHigh-Throughput Data

machine reading for cancer panomics - akbc · 2018. 11. 4. · genomics 12 discovery …...

Documents