Data Science & Bioinformatics
Advancing Medicine & Biology with Data Science
I am a Senior Bioinformatician at SickKids with a Ph.D. focused on developing predictive machine learning algorithms, scalable multi-omics pipelines, and clinical NLP tools that translate complex biological datasets into actionable discoveries.
99%+
Clinical NLP Accuracy
>80%
Manual Review Reduction
100+
Omics Datasets Analyzed
Featured Research
AI-Driven Injury Reporting in Pediatric Emergency Departments
JAMA Network Open
2025-07-01
Authors: Devin Singh, Alper Celik, Evangeline W J Zhang, Eric Liu, Daniel Rosenfield
Key Insight: NLP transformer models accurately identified 90% of pediatric injury cases from emergency department records, reducing manual chart review by over 80% and enabling real-time surveillance for CHIRPP.
Read Publication
Safety and Efficacy of Very Low-Dose Multi-Nut Oral Immunotherapy in Children
Clinical and Translational Allergy
2025-06-11
Authors: Julia EM Upton, Carmen H Li, Alireza Berenjy, Alana Galper, Xiaojun Yin, Alper Celik, Lucy Duan, Samantha Wong, Christina M Ditlof, Jennifer A Hoang, Moshe Ben-Shoshan, Akash Kothari, Lisa Hung, Mikhail Monteiro, Wut Hmone Phue, Theo J Moraes, Thomas Kaufmann, Thomas Eiwegger
Key Insight: Evaluates the safety and efficacy of very low-dose multi-nut oral immunotherapy for children with multiple nut allergies, demonstrating significant desensitization.
In Press / Coming Soon
Assessing the Diagnostic Impact of Blood Transcriptome Profiling in a Pediatric Cohort Previously Assessed by Genome Sequencing
NPJ Genomic Medicine
2025-06-11
Authors: Huayun Hou, Kyoko E. Yuki, Gregory Costain, Anna Szuto, Sierra Barnes, Arun K. Ramani, Alper Celik, Michael Braga, Meagan Gloven-Brown, Dimitri J. Stavropoulos, Sarah Bowdin, Ronald D. Cohn, Roberto Mendoza-Londono, Stephen W. Scherer, Michael Brudno, Christian R. Marshall, M. Stephen Meyn, Adam Shlien, James J. Dowling, Michael D. Wilson, Lianna Kyriakopoulou
Key Insight: Evaluated blood-derived RNA-seq in 134 pediatric patients with rare genetic disorders. RNA-seq reinforced or revised diagnoses in 36% of diagnosed cases and identified novel candidate genes in ~7% of previously undiagnosed cases.
Read Publication
Featured Projects
Benchmate
Benchmate is a data aggregation platform that enables scientists to fetch and integrate biological data from multiple sources including public APIs, literature databases, and molecular structure repositories. It provides specialized data structures...
CHIRPP AI & Clinical NLP Pipeline
The Canadian Hospitals Injury Reporting and Prevention Program (CHIRPP) tracks pediatric injury trends to inform public policy. We engineered a clinical NLP pipeline leveraging fine-tuned transformer models to extract structured diagnostic information...
Active Collaboration
Nut Allergy Biomarkers & Predictive Modeling
In collaboration with the Eiwegger Lab (University of Vienna & SickKids), this project develops statistical models to predict food allergy severity and treatment outcomes. We model allergen-induced intestinal contractions and immune responses...
Active Collaboration
Mnemosyne
Inspired by the Titan goddess of memory, Mnemosyne is an open-source suite designed to integrate qualitative research workflows. It streamlines data coding, synthesis, and knowledge extraction to make qualitative data analysis transparent,...
