- About Us
- Collaboration & Service
Constantin Aliferis MD, PhD, MS, Fellow of the American College of Medical Informatics
- Director, Biomedical Informatics Cores of NYU Clinical and Translational Science Institute and the NYU Cancer Center
- Director, Molecular Signatures Laboratory
- Scientific Director, Best Practices Integrative Informatics Consultation Service (BPIC)
- Scientific Director, NYULMC High Performance Computing Facility
- Associate Professor, Department of Pathology, NYU School of Medicine, Sackler Institute, NYU
- Adjunct Associate Professor of Biostatistics and Bioinformatics, Vanderbilt University.
Dr Aliferis has 17 years of formal graduate and post-graduate training in medicine, computer science (specialty area: intelligent systems/machine learning), and biomedical informatics. Following a 2-year appointment as Research Associate in Epidemiology in Athens University Medical School, he held faculty positions in Biomedical Informatics, Computer Science, Cancer Biology and Biostatistics at Vanderbilt University between 2000 and 2008. He was the Founding Director of the Vanderbilt MS/PhD Program in Biomedical Informatics in 2000-2001 and Founding Director of Vanderbilt’s Discovery Systems Lab between 2001 and 2008.
In 2008 Dr Aliferis joined NYU to develop and direct the NYU Center for Health Informatics and Bioinformatics (CHIBI), and holds appointments as Associate Professor of Pathology (with Adjunct Associate Professor appointments in Biomedical Informatics and Biostatistics at Vanderbilt, and with the Sackler Institute at NYU). Dr Aliferis is also Scientific Director of the Best Practices Informatics Consultation Core (BPIC) and of the High Performance Computing Facility.
Dr Aliferis has 80 peer reviewed full-length papers including in some of the most prestigious and highest impact factor journals in his fields of expertise (i.e., Machine Learning, Bioinformatics and Medical Informatics), as well as in leading biomedical journals, including: Journal of Machine Learning Research, Machine Learning Journal, Bioinformatics, BMC Bioinformatics, Journal of Computational Biology, JAMIA, Scientometrics, PLOS One, Cancer Research, Cell Host Microbe, PLoS Computational Biology, etc. Dr Aliferis has an additional 50 peer-reviewed publications including: 9 patents, 3 software systems, 1 published book and 3 more in contract to be published in 2011, and 5 peer-reviewed tutorials in national and international forums. Dr Aliferis has had 15 students (including 4 doctoral ones), and is mentoring 10 faculty members (4 as primary mentor). Several of his students earned national and international awards in informatics, including the Lindberg Fellowship, The Medinfo Gold Prize and the AMIA student paper competition first prize. Dr Aliferis has participated in 22 research projects (11 as PI or Director, 3 as co-PI, and 9 as Co-investigator) with funding provided by the NIH, NSF, the EU and industry in excess of $90 million in direct funding. Dr Aliferis leads the Informatics Core of the NYU CTSA and is also director of the NYU Cancer Center Biomedical Informatics Core.
Notable scientific contributions of Dr Aliferis and his group include:
(a) Dr Aliferis's lab has been a leading innovator in developing of local/global causal graph and Markov Blanket discovery algorithms used to discover pathways, biomarkers and construct molecular profiles. These include the first algorithms (IAMB family, 2002) that can discover Markov Blankets (for biomarker discovery and dimensionality reduction) correctly and with genomic-scale dimensionalities (prior algorithms were either incorrect/heuristic or non-scalable). He co-invented the first parallel distributed and chunked Markov Blanket algorithms (parallel IAMB, 2002), the first local causal neighborhood algorithms (HITON-PC and MMPC, 2002-3), the first correct, scaleable and sample efficient Markov Blanket algorithms (HITON and MMMB, 2003). Dr Aliferis lab also pioneered techniques that combined local with global causal graph learning (MMHC and LGL algorithm families, 2005-2009), generalized local learning (GLL 2008), combined SVM-Markov Blanket algorithms (FSMB, 2008), multiple Markov Blanket algorithms (TIE* family, 2008) and discovery of Markov Blankets in the presence of hidden variables (CIMB algorithms 2009).
(b) Dr Aliferis and his colleagues have developed, validated and disseminated software for the application of causal discovery methods and automated omics data analysis (Causal explorer, GEMS, FAST-AIMS) for data analysis and bioinformatics (with >2200 users in 50 countries covering most major universities and biomedical and computing companies).
(c) Dr Aliferis has conducted extensive benchmarking experiments for determining best practices in bioinformatics and has proved several theorems about the theoretical behavior of a variety of informatics algorithms.
(d) Dr Aliferis’ group introduced the application of machine learning methods to automatically assess and filter the quality and content of the biomedical literature and WWW health sites. In addition, his lab introduced the first models in the literature that correctly predict future number of citations for newly published articles within a subset of biomedicine and also they introduced the first models that differentiate between instrumental and non-instrumental citations and correct citation counts accordingly.
In terms of specific diseases and applications, Dr Aliferis and his collaborators and students have worked on molecular diagnostics for lung cancer, prediction of mortality due to community acquired pneumonia, mortality prediction in patients with acute respiratory distress syndrome, understanding physician decisions and guideline compliance in melanoma diagnosis and for neonatal sepsis diagnosis; they have also created ways to predict routine lab value abnormalities in order to reduce excessive resource utilization and lower the costs of care, and have created novel diagnostic models for stroke and stroke-like syndrome. Currently at NYU Dr Aliferis is working with colleagues to understand and force reduction of atherosclerotic plaques, understand the molecular basis of lung cancer, analyze microbiota sequence data in psoriasis, improve the treatment of keloids, personalize the treatment of mesothelioma and breast cancer, and unravel the genetic etiology of rheumatoid arthritis and breast cancer using GWAS data. Most of these projects have NIH and/or other federal funding.
Dr. Aliferis is also responsible for leading the design of the Research Enterprise Data Warehouse and research users of clinical data collected from the EPIC EMR/CPOE system.
227 E 30th Street, 7th Floor
New York University Langone Medical Center
New York, NY 10019, USA