| |
Use Cases: Charting BioPAX Pathways
The Pathway Toolkit is an open, web-based network of scriptable Java web apps that can display and manipulate XHTML CHARTSof
metadata on pathway models, cast into a context of related models in RAM that each each user can customize as desired
The toolkit network is deployed as two similar web apps working under the
client-server design pattern. The first is semi-public; the second private to one user:
- HTTP web site: a Pathway Warehouse Section
Each server instance is a web repository for files its sponsor wants to share with
people working locally or at related labs. It indexes
available resources, including BioPAX pathway models, bridge ontologies, and shared scripts for data I/O and processing.
Each client is similar but runs locally on one PC, customized as the
private agent for one end user. It can
interact with ALL available Warehouse sections to obtain or post copies
of resources, then use or perturb them further for specific R&D
purposes:
- WAR file download: Personal Pathway Toolkit
For
a reasonable license fee, any subscribing biomed group can
put a Warehouse Section on their intranet or the public
Internet, then download Personal versions to their staff, who gets access to all
source code. An industrial group should subscribe by contributing
cash to offset server hosting and toolkit R&D costs. Academics may optionally subscribe to a Personal version by
devoting worktime to BioPAX-support projects.
The Pathway Toolkit is built on top of the Lexikos WORDS modeler, a ontology editor for Topic Maps and NLP lexicons, with a range of general use cases.
Specializations of this baseline code for pathway vocabulary includes
the BioPAX ontology itself, a collection of shared modeling and data import/export tools, plus a
community-wide library of shared scripts for easily exploiting them.
Additional open source or custom tools can be separately contributed by
memebers of the community, optionally made scriptable via extensions to
the toolkit command language.
Other BioPAX-specific use cases are below. Most support Warehouse
sections, which often will be focused on
modeling a particular research context:.
1. Data Provenance
One
context (chart set) would portray the PRL
as a browseable semantic net giving info on
Suppliers and Pathway data bases. It would take periodic uploads
from today's PRL data base, and it would track and report recent
changes as "news" to the BioPAX community.
2. MetaModels
The toolkit optionally supports centralized indexing and search cycles for every Self-Annotating Identifier
whose triples get loaded into a context. This works with Data
Provenance tools to show who supplied what parts of what models, and
who else in the community or its contributors has published similar
triples.
A base of Ontology Design Patterns
supports the search tools above. By encouraging standard graph
structures for all of the important types of pathway elements, the
BioPAX ontology itself plays a key role in letting conformant data be
easily found, used and combined for any researchdomains that a given context may require.
3. VisualizationA
context can be built for any BioPAX pathways of interest, mixed with
other ontologies, then selectively viewed as pathway Charts under XSLT custom to each context. Final views will often be in HTML, optionally enhanced by custom CSS, Ajax.and or images
Typical near-term results may lack the elegance of high-end graphics like BioCyc, but our flexible charting approach does have unique advantages
- It offers ANY new view your context needs: XML, CSV, Lisp, Java graphics, etc
- It works even WHILE your contextual models are still actively evolving
- It thus may let you better visualize recent or needed changes to a context
Such approaches work both on single models and result sets from metamodel queries, so output page sets that resemble OwlDoc, yet eailsy evolve, will be common. Our design goal is to let each Pathway Toolkit be almost as easy to customize and edit as a wiki.
4. Converting Content into BioPAXOur toolkit lets example triples be incrementally posted by
suppliers, reviewed under uses cases 1-3, and aggregated with similar data well before
any bulk mapping is made to BioPAX exchange format. The step-by-step process
we advocate and support is simple, and it works for any data base - able to extract and (lightly) validate all relevant triples.
The Late 2005 Proposal
at BioPAX should quickly lead to statements of new triples that may be
relevant, but triples for 2.0 can already be input. Suppliers who
wish can export their data initially to the intermediate format of CSV
files, pending a later conversion by central BioPAX web
utilities. Check with us first on requirements,
Benefit #1: Data suppliers save
time and money by avoiding the need to learn exchange format details,
and sharing central conversion and reformatting tools. The might
cut work for early toolkit users by 3-4:1; and for later
users by 10:1
Benefit #2: Because suppliers work
easier and faster, data consumers more quickly get a large base of BioPAX pathway data in CVS format
they can use initially, plus the new tools below to help them exploit metadata formats where inferencing is actively supported
5. Import/Export Format ConversionsAs toolkit funding becomes available, we can expand its utility input forms
to convert posted triples
or CSV files directly into BioPAX 2.x data exchange files, fully
validated, and save them within a repository. Other import
formats can then be supported.
Initial export formats planned include standard BioPAX RDF, but also ISO-standard Topic Map files. Search, Visualization and change tracking toolkit
powers can them be applied to any imported file. Other export formats can then be added as convenient.
Benefit #1: Multiple Import formats, making it easier to merge
metadata across paradigms. Today we can handle (in early stages
at least), RDF, XTM, LTM, CVS, and WORDS scripts, a form of script specific to our toolkit that closely resembles Turtle.
Benefit #2:
Multiple future export formats of OWL-validated models make BioPAX more
widely useful, and easier to integrate with software designed around
other standards, paradigms or
languages, including CellML, Jesse, Prolog, Lisp, KIF, etc...
6. Integrated Pathway ModelerA
scriptable Pathway Toolkit, in eitrher of the above forms, can interact
not only with others like it across the web, but with lab tools,
ontologies, reasoners and data bases available only locally. All
of these custom local features must integrate at some level with
pathway models. This need leads directly to our Release Tools proposalSuch powers can greatly enhance the utility of a Pathway Warehouse, by arranging (with scripts) for pathway models that arise after their details are contextually perturbed. This typically will occur when edit scripts modify the loaded models of
typical BioPAX pathways, to carefully emulate the incremental changes expected from some
specific disease, drug, mutation, lab test, etc, then logically query the
predicted contextual results.
.
Big Benefit #1:
A Life Scientist can
first tap use cases 1-5 to select, merge, configure and incrementally
adjust a specific pathway variation to be explored - then download and
use its validated export semi-normally, enabling easy contrasts with standard-model results.
.
Big Benefit #2: Such a process can support any
other off-line tools or
processing models making sense to local Scientists. Mutated
BioPAX pathway models can be tested in ways that limit variations
systematically to all and only those of actual local interest.
|
