Transcript Slides

OpenBEL
Repurposable and computable knowledge
from biological findings extracted from the literature
SLA PHT Winter Virtual Conference • 07 December 2012
William Hayes, Ph.D., Senior VP Platform Development
Ted Slater, CTO OpenBEL Consortium
Selventa, Inc.
Mission
•
Librarians  Information Analysts
•
No longer sufficient to
•
Provide documents or books
•
Run searches
•
Literature is only valuable if you can use it
•
Must help to build and curate mission-critical databases
and knowledge bases
Purpose
•
Why is knowledge representation important?
•
What are issues that arise from using Excel as your only
database tech?
•
How does BEL fit?
“We are close to having
a $1,000
genome sequence,
but
this may be accompanied by
a
$1,000,000 interpretation.”
uce Korf, Former President of the American College of Medical Genetics
Knowledge
Representation
is important.
So, if you’re still doing it
the way you did 40 years
ago…
This is not
how
scientists
think about their
science.
This is.
Linguamatics I2E BEL
I2E Pilot for BEL
Causal Biology?
•
It’s likely that, to this day, you haven’t been able to buy or
develop a single knowledge base of causal biology you
could use effectively to
•
Understand disease biology
•
Do pathway walking
•
Predict, describe, and leverage mechanisms of action
and biomarkers competitively
Overview of
OPENBEL
What is OpenBEL?
•
The BEL Language
•
The BEL Tools
•
Biological Expression
Language
•
Open source (Apache
license)
•
Subject-predicate-object
triple-based, represents
causal relationships in
scientific findings
•
Knowledge acquisition &
management
•
Computable graph
knowledge bases from
BEL statements
•
APIs, other support
•
10 years of continuous
use and development at
Selventa
http://openbel.org
The BEL Language
BEL Highlights
•
•
BEL Terms
•
Represent types of abundances or processes
•
Based on external vocabularies and ontologies
•
Can be derived as functions of other terms
BEL Statements
•
Correspond to a specific finding
•
Express a relationship between terms
•
Retain context and provenance via annotations
BEL Statements
Term Expression
Relationship
Term Expression
Subject
Predicate
Object
decreases
a(CHEBI:corticosteroid) -| bp(NCI:”Tissue Damage”)
The abundance of molecules
designated by the name
“corticosteroid” in the CHEBI
namespace.
The biological process
designated by the name
“Tissue Damage” in the NCI
namespace.
Capture and Use Scientific
Findings
“RNA expression of RBL2 is directly mediated via
activation of the FOXO3 transcription factor”
Scientific Literature
tscript(p(HGNC:FOXO3)) => r(HGNC:RBL2)
J Biol Chem 2002 Nov 22 277(47) 45276-84
“LY294002 inhibits the activity of the
PI3K alpha catalytic subunit”
Original Research
a(PubChem:3973) -| kin(p(HGNC:PI3KCA))
XYZ Corp Document 12345
BEL Statement Annotations
Provide Context & Provenance
SET Citation = {“PubMed”, “J Mol Med (Berl). 2003
Mar;81(3):168-74.“ , “12682725”}
SET Evidence =
“high-dose steroid treatment decreases
vascular inflammation and ischemic tissue damage
after myocardial infarction and stroke through
direct vascular effects involving the
nontranscriptional activation of eNOS”
SET Tissue = “vascular system”
SET Disease = “Stroke”
a(CHEBI:corticosteroid) -| bp(NCI:”Tissue Damage”)
Assemble Facts as
Computable Models
BEL
statements…
…assemble
into models…
…for use in
applications
BEL and
“Reverse Causal Reasoning”
RCR – Using computable models
built from BEL sources to generate
mechanistic hypotheses based on
high-throughput data.
BEL Documents
RCR
Assembled Model
Hypothesis Network
The BEL Tools
The BEL Tools
•
A set of BEL-based applications and APIs that work
together to support
•
Knowledge acquisition and management
•
Knowledge publishing
•
Knowledge use
BEL Tools Components
•
BEL Framework v2.0 comprises
•
BEL Compiler/Assembler
•
•
Knowledge Assembly Models (KAMs)
•
•
Navigable networks (graphs) of biological facts
APIs
•
•
Creates KAMs from BEL documents
Java and Web Services Application Programming Interfaces for
interacting with KAMs stored in KAM Stores
Utilities
•
Validating and converting BEL documents; managing KAMs;
format conversions; etc.
•
Eclipse-based BEL Editor for editing BEL documents
•
BEL Cytoscape plug-in for network visualization
The OpenBEL Consortium
Promote, Protect, Standardize
•
Members of the OpenBEL Consortium proudly support
•
Driving adoption and evolving the standards
•
Increasing the amount of BEL content and number of BEL tools available to the
community
•
Increasing the number of people knowledgeable about OpenBEL through training,
education, and publication
•
Facilitating crucial administrative, technical, and legal functions for the
community to enable them to continue to advance the platform
•
Serving as a proxy between the academic community, the healthcare and life
sciences industry, publishers, and tools & services providers to ensure everyone’s
needs are best understood and that collaboration solutions are optimized
•
Holding neutral collaboration and legal events to ensure the platform is advanced
and protected
•
Serving as the neutral voice for OpenBEL to ensure that it remains a successful,
thriving, and growing platform
OpenBEL Ecosystem
BEL for BioPharma
•
Library services - capturing causal biology for use in R&D (disease
biology, MoA, Toxicology, Clinical Trials, Drug Safety)
•
Integrating databases (Ingenuity, AriadneGenomics, GeneGo,
internal research requests) into single KB
•
Model Organism Translatability - e.g. how to check if mouse biology
maps well with human biology for pre-clinical drug studies
•
Disease vs Normal Biology Comparisons - comparing known
biological differences between disease and normal states in an
organism's tissue
•
Assay Development - add tool compound and drug/target relations
into BEL Network to aid development of biological assays
Example using
KAM NAVIGATOR
Demonstration of
TRIPTYCH.JS
HTTP://WWW.TRIPTYCH.JS.COM/
Starting Questions
•
Does treating mouse model X with drug Y correlate to
human biology?
•
What is known biology for EGFR? Tissues expressed,
interacting proteins, associated disease conditions, …
•
What is found in the literature to associate kidney failure to
steatosis (liver adverse event)?
•
We need more detailed and highly contextualized
information for Pathway X than what is in
Ingenuity/GeneGo/etc
•
How could using drug X with MoA Y affect tissue Z?
Future Enhancements
•
BEL Statement/Script Creation
•
Dynamic Web-based Statement Editor (Q2 2013)
•
Curation Tool/PDF Reader Plugin (est Q4 2013)
•
Academic UIMA-based text mining workflows for KB
creation
Thank You!
Questions? Contact
[email protected] or visit http://openbel.org
Get OpenBEL at
https://github.org/OpenBEL