All Publications

Investigated whether the MAPT p.A152T variant increases susceptibility to Pick's disease, a rare 3R tauopathy, in 401 neuropathologically confirmed cases from the Pick's disease International Consortium. Found only one A152T carrier (0.12% frequency) versus 0.20% in healthy controls, and no significant differences in 4R tau expression via bulk RNA sequencing in carriers with Alzheimer's disease or progressive supranuclear palsy. Concluded that MAPT p.A152T does not increase Pick's disease risk and may instead associate with alternative tau pathologies.

Investigated the association of eight oxidative-stress-related SNPs (in GSTP1, SELENOS, CAT, SOD2, and EPHX1) with long COVID susceptibility in a clinically confirmed, bias-free cohort of 85 long COVID patients and 96 unvaccinated controls, all with mild or asymptomatic initial infection. Found associations between SOD2 and EPHX1 polymorphisms and disease progression, and released a 10-factor machine learning predictive model (balanced accuracy >0.74) as an open-access long COVID risk rating webtool.

Investigated the role of HLA allelic variants in neuropathological burden across 539 Lewy body disease cases using high-resolution whole-genome sequencing. Identified significant associations between specific HLA alleles and Lewy body counts, amyloid pathology, and disease subtype, highlighting immune-related genetic factors in LBD etiology.

Characterized MAPT alternative splicing patterns using long-read RNA sequencing in Pick's disease, a rare 3R tauopathy. Identified novel transcripts and complex splicing regulation extending beyond canonical isoforms, providing new insights into tau transcript diversity in the context of neurodegeneration.

Performed the first bulk transcriptomics study on Pick's disease using both short- and long-read sequencing on 28 PiD and 15 control parietal cortex samples. Identified CCL2 as the gene most strongly associated with 3R tau pathology and uncovered co-expressed gene networks involving mRNA processing, mitochondrial function, and immune processes.

Developed a machine learning-based web tool to predict COVID-19 severity from demographic and clinical features in 344 unvaccinated patients. Identified age, sex, hypertension, obesity, and cancer as key predictors of ICU admission, and deployed an optimized classification model as a publicly accessible prognostic tool.

Benchmarked hybrid and long-read genome assembly strategies using Illumina, Nanopore, and PacBio HiFi data on Drosophila and human genomes. Demonstrated that HiFi sequencing enables less computationally demanding assembly with novel algorithms, potentially democratizing genome assembly for smaller labs with limited resources.