research projects
- The mission of the PROOF Centre of
Excellence for the Prevention of Organ Failures is to develop
preventive, diagnostic or prognostic biomarkers to fight against heart,
lung and kidney failures. Some of the biomarker panels are currently being
validated in an international trials. I am the Chief Informatics Officer,
and lead a team
of statisticians and computer scientists to develop biomarkers by mining
transcriptomics, proteomics, metabolomics and clinical data. Apart from
publications in the medical venues, my team has published several papers
focusing on data cleansing, quality control and transformation for
transcriptomics and proteomics data. Currently there are two major projects
we are undertaking:
-
To develop a new biomarker-based blood test, HEARTBiT, for monitoring acute
cardiac transplant rejection in a laboratory environment that will expedite
translation of research to clinical value. Cardiac transplantation remains the
main intervention for those with end-stage heart failure. Maintenance
immunosuppression is given to all transplant recipients to prevent acute
rejection and loss of the allograft. Despite great improvements in
immunosuppressive therapies, acute rejection remains a clinical problem and
occurs at varying severity in 20-30% of patients within the first 3 months
post-transplant. Timely detection of moderate rejection allows for treatment
to be modified, preventing organ damage, graft failure and patient death. The
current method to monitor for rejection remains the endomyocardial biopsy
(EMB), a highly invasive and costly procedure that poses physical risks and
emotional stress to patients, who must undergo 12-15 such tests during the
first year post-transplant. EMB detects rejection only when tissue damage has
occurred, and lacks sensitivity as it provides information about tiny pieces
of the endomyocardium. Clearly, patients and clinicians would benefit from an
effective, cheaper, less invasive diagnostic test that can indicate when an
EMB is not needed.
-
To develop blood biomarker tests that will help clinicians identify who are
likely to have lung exacerbations within 3 months and subclassify exacerbations to
enable targeted therapy. Chronic obstructive pulmonary disease (COPD) is a
common inflammatory disorder of the lung, characterized by shortness of breath
and progressive loss in lung function. It affects 2.6 million Canadians and is
the leading cause of hospitalizations in Canada, responsible for $12
billion/year in direct health care expenditures. Over the next 15 years, the
number of Canadians with COPD will more than double and the number of hospital
beds required to manage them will triple. Our vision is to harness the power
of multi-omics and machine learning methods to advance biomarker solutions
that will enable precision health in COPD, relieve patient suffering, and
reduce the enormous socioeconomic burden of COPD in Canada.
- BIN was the Business
Intelligence Network funded by NSERC, IBM and SAP. Business
Intelligence is the commercial term for using information within an
organization to make informed decisions and to run operations effectively
based on known data. As a research group, this network conducted research
revolving around (i) policy and strategy management; (ii) capitalizing on
document assests; (iii) adaptive data cleaning; and (iv) business-driven
data integration. The network was funded till 2014. I was the
associate director of the network. My projects within this network continue
and focus on text mining and natural language processing. Specifically, together with Giuseppe Carenini, we have done extensive research
on feature extraction
and conversation identification from informal documents such as emails,
blogs, and meeting notes. We produce extractive and abstractive summaries
of the analysis in natural language, in information visualization style, or
a carefully crafted combination of the two. More details can be found in
the Natural Language Processing Research group page and the
book entitled "Methods
for Mining and Summarizing Text Conversations".
- MERIDIAN:
Marine Environmental Research Infrastructure for Data Integration and
Application Network is a project funded by CFI. Ocean development must be
done sustainably, which includes controlling and/or mitigating noise impacts.
Noise in the ocean from shipping and other offshore industrial activity has a
significant impact on protected marine species. Acoustic ocean data is currently
collected in non-structured and separate databases. The MERIDIAN Consortium,
including 12 Canadian research organizations and global world-class ocean
acoustic researchers, will develop a research data infrastructure to consolidate
and support ocean acoustic data. This infrastructure will be a widely used
resource for the Canadian and international academic community to drive
discovery and innovation while supporting the development of the ocean
soundscape atlas for Canada. By focusing on acoustic data, MERIDIAN will enhance
Canada's global leadership in the field. No other country has established a
national data resource for noise in the ocean. MERIDIAN will create a cloud-
based platform of research tools based on data science methods, techniques and
tools. The infrastructure will support data discovery, data integration and
interoperability, interactive data visualization, and data analysis for
streaming data. MERIDIAN will enable Canada's leading ocean researchers to fully
exploit Canada's ocean data, monitoring trends in the state of the ocean
acoustic environment, and enable more timely, effective and efficient protection
of valued marine species and protected areas.
Last updated: 28/10/2018
