Chan Zuckerberg Initiative

A masked language model for T-cell receptor and peptide-MHC interaction prediction using compositional pretraining and entropy-guided non-autoregressive decoding.

A fine-tuned scLDM variant trained on 14.5 million CD4+ T cells for counterfactual prediction of single-gene perturbation effects in immune cells.

A scalable latent diffusion model for generating realistic single-cell gene expression profiles, using a permutation-invariant VAE and flow-matching diffusion transformer.

A 1.2-billion-parameter hierarchical transformer that predicts personalized gene expression from diploid genomes, integrating individual genetic variants for ancestry-robust eQTL analysis.

A reasoning language model post-trained on virtual cell simulations to answer complex biological questions about gene perturbations in natural language.

A graph-signal-processing foundation model that embeds gene regulatory network structure directly into its attention mechanism for parameter-efficient single-cell transcriptomics.

A generative cross-species foundation model for single-cell transcriptomics, trained on 112 million cells from 12 species spanning 1.5 billion years of evolution.

Eighth-place CZII CryoET Kaggle solution; a weighted model soup of tiny, medium, and large 3D U-Nets pretrained on simulated data and fine-tuned on experimental cryo-ET tomograms.

First-place solution from the CZII CryoET Object Identification Kaggle competition; an ensemble of 3D EfficientNet-encoder U-Nets for multi-class protein particle picking.

Fifth-place solution from the CZII CryoET Kaggle competition; an ensemble of four lightweight 3D U-Nets for protein particle localization in cryo-ET tomograms.

Seventh-place CZII CryoET Kaggle solution; an ensemble of three heatmap-predicting 3D segmentation models using ResNet50d and EfficientNetV2-M backbones for particle picking.

A deep learning framework for multi-class protein particle picking in cryo-ET tomograms using a 3D U-Net with automated architecture search via Bayesian optimization.

A SAM2-based deep learning framework for vesicle segmentation in cryo-ET tomograms and 2D micrographs, supporting both zero-shot and fine-tuned inference.

A variational autoencoder for interpretable 3D reconstruction and representation learning of protein subtomograms from cryo-ET data, trained on 5.8 million synthetic particles.

A differentiable wave-optical framework for label-agnostic computational microscopy of biomolecular density and orientation across diverse imaging modalities.

Self-supervised Vision Transformer models trained on proteome-wide fluorescence microscopy images from the Human Protein Atlas for subcellular protein localization.

A single-cell perturbation model that augments scGPT with gene-level language embeddings from NCBI, UniProt, and Gene Ontology to improve multi-gene perturbation prediction.

A unified diffusion model enabling bidirectional transformation between protein amino acid sequences and fluorescence microscopy images for subcellular localization prediction.

A self-supervised contrastive learning method for embedding cell and organelle dynamics from time-lapse microscopy using temporal regularization and single-cell tracking.

A variational autoencoder pretrained on 74 million human single-cell transcriptomes from the CELLxGENE Census for scalable batch correction, cell typing, and data integration.