Genomics
Spatial & single-cell
Spatial transcriptomics and scRNA-seq to resolve where — and in which cells — interferon signaling remodels the lung.
Research
Immune recognition, and what happens when it misfires.
The immune system’s first job is to tell danger from self. When that judgment fails in the lung, the result is inflammation the body inflicts on itself. That failure — its molecular triggers and its tissue consequences — is the throughline of this work.
Current focus · Shum Lab, UCSF
Immune-driven lung disease
When an internal alarm won’t switch off, the lung pays for it.
Fig 01 · Shum LabSTING trafficking, derailed through the COPI coatomer, floods the lung with type I interferon.
In the laboratory of Anthony Shum at UCSF, the focus is COPA syndrome and a family of related disorders in which the innate immune sensor STING becomes inappropriately active — flooding the lung with type I interferon and driving interstitial lung disease.
Using spatial transcriptomics and single-cell RNA sequencing, the work maps how disruptions in the COPI coatomer machinery derail the trafficking of STING inside the cell, and how those molecular missteps reshape lung tissue — cell type by cell type, region by region.
These conditions are rare. Their lessons are not: the same interferon biology reaches into far more common forms of pulmonary fibrosis and autoimmunity.
A rare disease is often the clearest window onto a common one.
Doctoral work · Desrosiers Lab
Disarming HIV’s shield
The mirror image of autoimmunity: a virus that hides too well.
Fig 02 · Desrosiers LabAn O-linked glycan shield hides HIV-1 envelope from a class of broadly neutralizing antibodies.
Doctoral research examined the virologic determinants of elite control of HIV — the rare patients whose immune systems hold the virus in check without treatment.
The central finding: certain HIV-1 strains elongate a region of their envelope protein and cloak it in O-linked sugars — a glycan shield that hides the virus from a class of broadly neutralizing antibodies. The result helps explain how the virus evades immune recognition, and informs how next-generation antibody therapies might be designed to overcome it.
To design a better antibody, first understand how the virus learned to hide.
Approach
Methods across scales
From single molecules to whole tissues — pairing wet-lab immunology with genomics, structural biology, and computation to read the same problem at every resolution.
Structure
Trafficking machinery
Structural and computational modeling of the COPI coatomer and STING to understand how transport goes wrong.
Bench & bedside
Translational lens
Rare monogenic disease as a controlled experiment in nature, read back onto common fibrotic and autoimmune lung disease.
Next