View
27
Download
1
Category
Tags:
Preview:
Citation preview
Centro de Investigación ProS
Applying Capability Modelling in the Genomics Diagnosis Domain: Lessons Learned
Francisco Valverde and Maria José Villanueva
2nd International Workshop on Capability-oriented Business Informatics (CoBI 2015). 16th of July, 2015
Agenda
1. Motivation
2. Applying CDD in a genomics SME
3. Lessons Learned
4. Conclusions
3
Motivation
Context
Innovative services for Digital Enterprises with ORCA Capability as a Service for Digital Enterprises
Apply CaaS results into innovative case studies from a Spanish region
4
Motivation
Case study: Geneticists from a SME (IMEGEN) provide their disease diagnosis services using genetic information
They provide a portfolio of genetic tests to be carried out (around 1.000 different tests)
6
Motivation
Genetics is a continuously evolving context:✘There is no an standard software solution because easily
will become outdated✘Lack of Software Engineering / Conceptual modelling
practices for supporting evolution✘Need of novel and powerful infrastructure
… but the underlying process remains the same
8
Motivation
Impact in Geneticists’ work:
Addressing tedious programming tasks to customize tools
Spending more time learning computer science issues
Making mistakes due to lack of knowledge and manual
procedures.
9
Motivation
Impact in business
Could you provide a genetic test for the novel disease X?
Predict expected delivery time for a test
Compliance with new laboratory ISO regulations
Reducing costs by infrastructure outsourcing (Cloud
technologies or external provider)
They don’t really know if they can provide their capabilities in
the near future!!!
10
Motivation
This works deals with this issue from two
perspectives
Business: using CDD to formalize the genetic
diagnosis capability they must provide
Technical: analyzing Bioinformatics Workflow
Management Systems (BWMS) to support the
capability deployment
11
Applying CDD
According to CaaS Project: the ability and capacity that enables an enterprise to achieve a business goal in a certain operational context
The goal to accomplish
The ability to engineer a bridge
The capacity such as money or tools to build a bridge
The context in which the bridge must be build
(location)
12
Applying CDD
Why CDD?• Enterprise context clearly affects the service delivery in
this use case• Know-how reuse is feasible in the domain as pipeline (data
processing workflows)• Lack of widely accepted conceptual models / standards to
express genetic data
In our view, CDD addresses these three main concerns using a sound approach
13
Applying CDD
We interviewed with 3 geneticists from the SME to:• Define a domain model as a conceptual schema• Understand goals, KPIs and current bioinformatics context• Formalize their current process model• Understand the technological tools involved in the process• Detect current bottlenecks
We specified its business as capabilities following a
template
14
Applying CDD
Capability template (from CaaS)
Goal: Desired state of affairs that needs to be attained.
Goal KPI: KPI) or monitoring the achievement of a goal.
Context: Information characterizing the situation in which a
business capability should be provided.
Capacity: Availability of resources for delivering the capability
Ability: Level of available competence of a enterprise to
accomplish a goal.
15
Applying CDD
Provide a disease genomic Diagnosis (Main capability)
Goal: Provide a accurate diagnosis regarding a genomic
disease
Capacity: NGS machine and technological infrastructure
(Server, Disk Array etc.)
Ability: geneticists with knowledge about data sources
with trustful information and the genomic diagnosis
process
16
Applying CDD
Overall Process
17
Applying CDD
IMEGEN Process
18
Applying CDD
This capability is easy to manage as four sub-capabilities:
1. Provide a Genomic Diagnosis
1.1 Provide integrated information from public data source
1.2 Support novel bioinfomatics services
1.3 Management of new genomic data
1.4 End-user (friendly reports physician)
19
Applying CDD
1. Provide integrated information from public data sources • Context: New datasets to be included and updated
versions• KPI: number of supported datasets
2. Support novel bioinformatics services• Context: New algorithms, new sequencing technologies,
data processing utilities, novel IS architectures• KPI: number of supported services, response-time
20
Applying CDD
3. Management of new genomic data• Context: New discoveries about genomic mechanisms and
disease• KPI: disease knowledge
4. End-user (friendly reports physician) • Context: New laws and standards (ISO) regarding clinic
analyses• KPI: Law/Certification compliance and trust
21
Applying CDD
Bioinformatics Workflow Management Systems:
describe workflows made up of software
components that manipulate genetic data.
End-user oriented: they provide some guidance for
creating experiments, such as visual notations or
wizards
Three analyzed: Taverna, Galaxy, e-bioflow
22
Applying CDD
Example of a BWMS (Galaxy)
23
Applying CDD
Support for each CapabilityCapability Taverna Galaxy eBioFlow
Integrated information Partial Partial Partial
Support novel bioinfomatics services Yes Partial No
Data management Partial No Partial
End-user friendly reports Partial Partial No
24
Lessons Learned
Regarding BWMS:• Taverna is the most complete but does not take into
account domain knowledge• Galaxy, simpler workflow notation and provides a lot of
functionality out-of-the box• eBioflow, provides a good workflow language in terms of
expressivity and a user-friendly interface but lacks of advanced functionality
Galaxy was selected because of the advanced functionality provided
25
Lessons Learned
Geneticists state that capabilities specification are a
nice and organized documentation of their process.
CDD overcomes the worfklow-oriented vision in
bioinformatics
26
Lessons Learned
CDD proposed conceptual models are useful for data
information retrieval and domain modelling
Re-using of know-how to address novel genetic
diseases
CDD + BWMS offers a clear improvement over
current practices and future evolution
27
Conclusions
We have present the potential benefits of applying
CDD in a novel domain
Problem specification (Capability) is decoupled from
implementation (BWMS)
As further work we will evaluate in practice the
analyzed capabilities using Galaxy
28
Questions/Comments
{fvalverde, mvillanueva}@pros.upv.es
Recommended