Erik Swanson

Latent generative search unlocks de novo design of untapped biomolecular interactions at scale

Reward-guided inference-time search over the Proteína-Complexa sequence-structure codesign model yields validated binders across 127 targets, kinases and a free carbohydrate, with codesigned sequences outperforming ProteinMPNN redesign.

Spatial Transcriptomics Reveals Compartment-Specific Immune Activation Signatures in Ileal and Lymph Node Tissue in Treated HIV Infection

This study uses spatial transcriptomics to identify compartment-specific immune activation signatures in ileal and lymph node tissues of HIV-treated patients, revealing tissue-specific transcriptional differences linked to immune activation and microbial translocation.

Plasma metabolomics reveals lipid-predominant metabolic disruptions in virologically suppressed people with HIV

The paper uses metabolomics to show that HIV-treated individuals have metabolic disruptions, including lipid changes and inflammation markers, despite viral suppression.

Multi-omics links microbial dysbiosis, systemic inflammation and metabolomic disruptions to SNAE risk in treated HIV

This study links microbial dysbiosis, systemic inflammation, and metabolomic disruptions to SNAE risk in treated HIV patients through multi-omic analysis.

Gut barrier integrity biomarkers are associated with increased inflammation and predict disease status in hospitalized COVID-19 patients

This study identifies gut barrier integrity biomarkers linked to inflammation and disease severity in hospitalized COVID-19 patients using plasma biomarker analysis and a random forest model.

Pre-treatment Microbiome Diversity and Function is associated with Expansion of Cytotoxic and Regulatory Immune Populations after N-803 treatment in People with HIV

This study shows that pre-treatment gut microbiome diversity in HIV patients correlates with immune activation and reduced HIV RNA+ cells after N-803 treatment.

mBER: Controllable de novo antibody design with million-scale experimental screening

The paper introduces mBER, a system for designing antibody-format binders with high experimental success rates, achieving 45% success across 145 targets and 38% binding rates to specific epitopes.