Physics-driven structural docking and protein language models accelerate antibody screening and design for broad-spectrum antiviral therapy.

Therapeutic antibodies have become one of the most influential therapeutics in modern medicine to fight against infectious pathogens, cancer, and many other diseases. However, experimental screening for highly efficacious targeting antibodies is labor-intensive and of high cost, which is exacerbated by evolving antigen targets under selective pressure such as fast-mutating viral variants. As a proof-of-concept, we developed a machine learning-assisted antibody generation pipeline that greatly accelerates the screening and re-design of immunoglobulins G (IgGs) against a broad spectrum of SARS-CoV-2 coronavirus variant strains. These viruses infect human host cells via the viral spike protein binding to the host cell receptor angiotensin-converting enzyme 2 (ACE2). Using over 1300 IgG sequences derived from convalescent patient B cells that bind with spike's receptor binding domain (RBD), we first established protein structural docking models in assessing the RBD-IgG-ACE2 interaction interfaces and predicting the virus-neutralizing activity of each IgG with a confidence score. Additionally, employing Gaussian process regression (also known as Kriging) in a latent space of an antibody language model, we predicted the landscape of IgGs' activity profiles against individual coronaviral variants of concern. With functional analyses and experimental validations, we efficiently prioritized IgG candidates for neutralizing a broad spectrum of viral variants (wildtype, Delta, and Omicron) to prevent the infection of host cells in vitro and hACE2 transgenic mice in vivo. Furthermore, the computational analyses enabled rational redesigns of selective IgG clones with single amino acid substitutions at the RBD-binding interface to improve the IgG blockade efficacy for one of the severe, therapy-resistant strains - Delta (B.1.617). Our work expedites applications of artificial intelligence in antibody screening and re-design even in low-data regimes combining protein language models and Kriging for antibody sequence analysis, activity prediction, and efficacy improvement, in synergy with physics-driven protein docking models for antibody-antigen interface structure analyses and functional optimization.

Epigenetic dysregulation in Alzheimer's disease peripheral immunity.

The peripheral immune system in Alzheimer's disease (AD) has not been thoroughly studied with modern sequencing methods. To investigate epigenetic and transcriptional alterations to the AD peripheral immune system, we used single-cell sequencing strategies, including assay for transposase-accessible chromatin and RNA sequencing. We reveal a striking amount of open chromatin in peripheral immune cells in AD. In CD8 T cells, we uncover a cis-regulatory DNA element co-accessible with the CXC motif chemokine receptor 3 gene promoter. In monocytes, we identify a novel AD-specific RELA transcription factor binding site adjacent to an open chromatin region in the nuclear factor kappa B subunit 2 gene. We also demonstrate apolipoprotein E genotype-dependent epigenetic changes in monocytes. Surprisingly, we also identify differentially accessible chromatin regions in genes associated with sporadic AD risk. Our findings provide novel insights into the complex relationship between epigenetics and genetic risk factors in AD peripheral immunity.

Simultaneous CRISPR screening and spatial transcriptomics reveals intracellular, intercellular, and functional transcriptional circuits.

Pooled optical screens have enabled the study of cellular interactions, morphology, or dynamics at massive scale, but have not yet leveraged the power of highly-plexed single-cell resolved transcriptomic readouts to inform molecular pathways. Here, we present Perturb-FISH, which bridges these approaches by combining imaging spatial transcriptomics with parallel optical detection of in situ amplified guide RNAs. We show that Perturb-FISH recovers intracellular effects that are consistent with Perturb-seq results in a screen of lipopolysaccharide response in cultured monocytes, and uncover new intercellular and density-dependent regulation of the innate immune response. We further pair Perturb-FISH with a functional readout in a screen of autism spectrum disorder risk genes, showing common calcium activity phenotypes in induced pluripotent stem cell derived astrocytes and their associated genetic interactions and dysregulated molecular pathways. Perturb-FISH is thus a generally applicable method for studying the genetic and molecular associations of spatial and functional biology at single-cell resolution.

Scalable genetic screening for regulatory circuits using compressed Perturb-seq.

Pooled CRISPR screens with single-cell RNA sequencing readout (Perturb-seq) have emerged as a key technique in functional genomics, but they are limited in scale by cost and combinatorial complexity. In this study, we modified the design of Perturb-seq by incorporating algorithms applied to random, low-dimensional observations. Compressed Perturb-seq measures multiple random perturbations per cell or multiple cells per droplet and computationally decompresses these measurements by leveraging the sparse structure of regulatory circuits. Applied to 598 genes in the immune response to bacterial lipopolysaccharide, compressed Perturb-seq achieves the same accuracy as conventional Perturb-seq with an order of magnitude cost reduction and greater power to learn genetic interactions. We identified known and novel regulators of immune responses and uncovered evolutionarily constrained genes with downstream targets enriched for immune disease heritability, including many missed by existing genome-wide association studies. Our framework enables new scales of interrogation for a foundational method in functional genomics.

Compressed Perturb-seq: highly efficient screens for regulatory circuits using random composite perturbations.

Pooled CRISPR screens with single-cell RNA-seq readout (Perturb-seq) have emerged as a key technique in functional genomics, but are limited in scale by cost and combinatorial complexity. Here, we reimagine Perturb-seq's design through the lens of algorithms applied to random, low-dimensional observations. We present compressed Perturb-seq, which measures multiple random perturbations per cell or multiple cells per droplet and computationally decompresses these measurements by leveraging the sparse structure of regulatory circuits. Applied to 598 genes in the immune response to bacterial lipopolysaccharide, compressed Perturb-seq achieves the same accuracy as conventional Perturb-seq at 4 to 20-fold reduced cost, with greater power to learn genetic interactions. We identify known and novel regulators of immune responses and uncover evolutionarily constrained genes with downstream targets enriched for immune disease heritability, including many missed by existing GWAS or trans-eQTL studies. Our framework enables new scales of interrogation for a foundational method in functional genomics.

Cerebrospinal fluid immune dysregulation during healthy brain aging and cognitive impairment.

Cerebrospinal fluid (CSF) contains a tightly regulated immune system. However, knowledge is lacking about how CSF immunity is altered with aging or neurodegenerative disease. Here, we performed single-cell RNA sequencing on CSF from 45 cognitively normal subjects ranging from 54 to 82 years old. We uncovered an upregulation of lipid transport genes in monocytes with age. We then compared this cohort with 14 cognitively impaired subjects. In cognitively impaired subjects, downregulation of lipid transport genes in monocytes occurred concomitantly with altered cytokine signaling to CD8 T cells. Clonal CD8 T effector memory cells upregulated C-X-C motif chemokine receptor 6 (CXCR6) in cognitively impaired subjects. The CXCR6 ligand, C-X-C motif chemokine ligand 16 (CXCL16), was elevated in the CSF of cognitively impaired subjects, suggesting CXCL16-CXCR6 signaling as a mechanism for antigen-specific T cell entry into the brain. Cumulatively, these results reveal cerebrospinal fluid immune dysregulation during healthy brain aging and cognitive impairment.

Compressed sensing for highly efficient imaging transcriptomics.

Recent methods for spatial imaging of tissue samples can identify up to ~100 individual proteins1-3 or RNAs4-10 at single-cell resolution. However, the number of proteins or genes that can be studied in these approaches is limited by long imaging times. Here we introduce Composite In Situ Imaging (CISI), a method that leverages structure in gene expression across both cells and tissues to limit the number of imaging cycles needed to obtain spatially resolved gene expression maps. CISI defines gene modules that can be detected using composite measurements from imaging probes for subsets of genes. The data are then decompressed to recover expression values for individual genes. CISI further reduces imaging time by not relying on spot-level resolution, enabling lower magnification acquisition, and is overall about 500-fold more efficient than current methods. Applying CISI to 12 mouse brain sections, we accurately recovered the spatial abundance of 37 individual genes from 11 composite measurements covering 180 mm2 and 476,276 cells.

Using viral load and epidemic dynamics to optimize pooled testing in resource-constrained settings.

Virological testing is central to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) containment, but many settings face severe limitations on testing. Group testing offers a way to increase throughput by testing pools of combined samples; however, most proposed designs have not yet addressed key concerns over sensitivity loss and implementation feasibility. Here, we combined a mathematical model of epidemic spread and empirically derived viral kinetics for SARS-CoV-2 infections to identify pooling designs that are robust to changes in prevalence and to ratify sensitivity losses against the time course of individual infections. We show that prevalence can be accurately estimated across a broad range, from 0.02 to 20%, using only a few dozen pooled tests and using up to 400 times fewer tests than would be needed for individual identification. We then exhaustively evaluated the ability of different pooling designs to maximize the number of detected infections under various resource constraints, finding that simple pooling designs can identify up to 20 times as many true positives as individual testing with a given budget. Crucially, we confirmed that our theoretical results can be translated into practice using pooled human nasopharyngeal specimens by accurately estimating a 1% prevalence among 2304 samples using only 48 tests and through pooled sample identification in a panel of 960 samples. Our results show that accounting for variation in sampled viral loads provides a nuanced picture of how pooling affects sensitivity to detect infections. Using simple, practical group testing designs can vastly increase surveillance capabilities in resource-limited settings.

Using viral load and epidemic dynamics to optimize pooled testing in resource constrained settings.

Extensive virological testing is central to SARS-CoV-2 containment, but many settings face severe limitations on testing. Group testing offers a way to increase throughput by testing pools of combined samples; however, most proposed designs have not yet addressed key concerns over sensitivity loss and implementation feasibility. Here, we combine a mathematical model of epidemic spread and empirically derived viral kinetics for SARS-CoV-2 infections to identify pooling designs that are robust to changes in prevalence, and to ratify losses in sensitivity against the time course of individual infections. Using this framework, we show that prevalence can be accurately estimated across four orders of magnitude using only a few dozen pooled tests without the need for individual identification. We then exhaustively evaluate the ability of different pooling designs to maximize the number of detected infections under various resource constraints, finding that simple pooling designs can identify up to 20 times as many positives compared to individual testing with a given budget. We illustrate how pooling affects sensitivity and overall detection capacity during an epidemic and on each day post infection, finding that sensitivity loss is mainly attributed to individuals sampled at the end of infection when detection for public health containment has minimal benefit. Crucially, we confirm that our theoretical results can be accurately translated into practice using pooled human nasopharyngeal specimens. Our results show that accounting for variation in sampled viral loads provides a nuanced picture of how pooling affects sensitivity to detect epidemiologically relevant infections. Using simple, practical group testing designs can vastly increase surveillance capabilities in resource-limited settings.