Biology subjects

Ricci-Tam, C.

Publications and source records attributed to Ricci-Tam, C..

11 recordsLinked to original sources

bioRxiv2026

Systematic mapping of the prokaryotic noncoding sequence space with a genomic language model

Prokaryotic genomes encode a vast repertoire of noncoding elements such as ribozymes, riboswitches, and CRISPR RNAs. The sequence-level diversity of these intergenic elements precludes alignment-based clustering that would enable annotation and discovery of novel noncoding RNA families. Here, we show that a genomic language model fine-tuned on prokaryotic intergenic sequences learns representations of noncoding elements from DNA sequence. We used these embeddings to organize 376 million intergenic regions from 1.3 terabase pairs of prokaryotic genomes and metagenomes into 5.14 million clusters. This approach nominated 20,149 clusters of putative structured RNAs with a candidate protein partner. Periodicity in token entropies revealed internal architecture, recovering CRISPR, TIGR and VIPR arrays together with their cognate effector proteins from sequence alone. Targeted clustering of bridge recombinase-adjacent sequences identified bridge RNA variants with longer guide recognition sequences than any characterized ortholog. Finally, we directed teams of AI agents to mine the database under researcher-defined criteria, uncovering divergent riboswitches, ribozymes, and other structured RNAs with biological hypotheses for their function. We release a publicly available database that transforms the dark matter of prokaryotic genomes into an explorable resource for human-agent molecular discovery.

genomics↗

bioRxiv2026

Spatially resolved, multimodal in vivo Perturb-seq using antibody-based cell hashing

Large-scale perturbation screens have begun to map cell intrinsic gene-function relationships, yet how genes shape tissue architecture remains largely unexplored. To address this gap, we developed PerturbSpace, a novel approach that integrates CRISPR perturbations with spatially hashed single-cell multiomics. This approach enables the first high-throughput, spatially resolved Perturb-seq analysis across complex tissue architecture in vivo. Notably, PerturbSpace enables spatial transcriptome-wide perturbation readouts at organ scale and can be seamlessly integrated with orthogonal modalities. We combine PerturbSpace with surface proteomics and expressed lineage tracing barcodes to demonstrate multimodal compatibility. We use PerturbSpace to study the genetic determinants of tissue architecture. First, we map how 40 transcriptional regulators determine the size and lineage composition of colonies in the spleen during regenerative hematopoiesis. Second, we characterize immune-niche interactions in the liver by dissecting the extrinsic effects mediated by cytokine-secreting immune cells on their neighboring cells. Collectively, our work establishes PerturbSpace as a scalable and cost-effective approach for transcriptome-wide spatial profiling of whole cells while remaining compatible with the single-cell multiomics workflows that the field has already adopted at scale.

genomics↗

bioRxiv2026

BioReason-Pro: Advancing Protein Function Prediction with Multimodal Biological Reasoning

Protein function annotation is fundamental to understanding biological mechanisms, designing therapeutics, and advancing biomedical research. Current computational methods either rely on shallow sequence similarity or treat function prediction as isolated classification tasks, failing to capture the integrative reasoning across sequence, structure, domains, and interactions that expert biologists perform to infer function. We introduce BioReason-Pro, the first multimodal reasoning large language model (LLM) for protein function prediction that integrates protein embeddings with biological context to generate structured reasoning traces. A key input into BioReason-Pro is the set of GO term predictions made by GO-GPT, our autoregressive transformer that captures hierarchical and cross-aspect dependencies of GO terms. BioReason-Pro is trained via supervised fine-tuning on synthetic reasoning traces generated by GPT-5 for over 130K proteins and further optimized through reinforcement learning. It achieves 73.6% Fmax on GO term prediction and an LLM judge score of 8/10 on functional summaries, substantially outperforming previous methods. Evaluations with human protein experts show that BioReason-Pro annotations are preferred over ground truth UniProt annotations in 79% of cases. Remarkably, BioReason-Pro predicted a novel interaction partner for the renal cancer biomarker RCDG1, which we confirmed in the lab by co-immunoprecipitation. In other binding-partner predictions, its per-residue attention localized to the exact contact residues resolved in cryo-EM structures. Together, GO-GPT and BioReason-Pro establish a framework for protein function prediction that combines precise ontology modeling with interpretable biological reasoning.

molecular biology↗

bioRxiv2026

Stack: In-Context Learning of Single-Cell Biology

Foundation models trained on single-cell transcriptomic data offer the promise of identifying and predicting the diversity of cellular phenotypes across species, diseases, and other biological conditions. However, the current models are limited to their supervised training conditions and tasks, which limits their utility for biological discovery. Here, we present SO_SCPLOWTACKC_SCPLOW, a foundation model trained on 149 million uniformly preprocessed human single cells that leverages tabular attention to generate representations for each cell informed by the cells in its context. SO_SCPLOWTACKC_SCPLOW offers substantial improvements for downstream tasks in the zero-shot setting compared to baselines, whether they are zero-shot, fine-tuned, or trained from scratch on the target dataset. SO_SCPLOWTACKC_SCPLOW can perform in-context learning from unlabeled cells representing arbitrary conditions, such as a chemical perturbation or a different donor, and predict the effect of those conditions on a target cell population without requiring data-specific fine-tuning. We apply SO_SCPLOWTACKC_SCPLOW to generate Perturb Sapiens, the first human whole-organism atlas of perturbed cells, spanning 28 tissues, 40 cell types, and 892 drug, cytokine, and genetic perturbations. We validated subsets of Perturb Sapiens using in vitro stimulation profiles. SO_SCPLOWTACKC_SCPLOW uniquely empowers prioritization of donor-specific perturbation effects, a capability we validated in our newly collected DiseasePert-3M data, comprising T cells from 40 donors across 14 diseases, stimulated with 11 cytokines. Overall, SO_SCPLOWTACKC_SCPLOW presents a new modeling framework where cells themselves act as guiding examples at inference time, unlocking general-purpose in-context learning capabilities for single-cell biology.

bioinformatics↗

bioRxiv2025

Predicting cellular responses to perturbation across diverse contexts with STATE

Cellular responses to perturbations are a cornerstone for understanding biological mechanisms and selecting drug targets. While machine learning models offer tremendous potential for predicting perturbation effects, they currently struggle to generalize to unobserved cellular contexts. Here, we introduce SO_SCPLOWTATEC_SCPLOW, a transformer model that predicts perturbation effects while accounting for cellular heterogeneity within and across experiments. SO_SCPLOWTATEC_SCPLOW predicts perturbation effects across sets of cells and is trained using gene expression data from over 100 million perturbed cells. SO_SCPLOWTATEC_SCPLOW improved discrimination of effects on large datasets by more than 30% and identified differentially expressed genes across genetic, signaling and chemical perturbations with significantly improved accuracy. Using its cell embedding trained on observational data from 167 million cells, SO_SCPLOWTATEC_SCPLOW identified strong perturbations in novel cellular contexts where no perturbations were observed during training. We further introduce Cell-Eval, a comprehensive evaluation framework that highlights SO_SCPLOWTATEC_SCPLOWs ability to detect cell type-specific perturbation responses, such as cell survival. Overall, the performance and flexibility of SO_SCPLOWTATEC_SCPLOW sets the stage for scaling the development of virtual cell models.

systems biology↗

bioRxiv2025

Megabase-scale human genome rearrangement with programmable bridge recombinases

Bridge recombinases are a class of naturally occurring RNA-guided DNA recombinases. We previously demonstrated they can programmably insert, excise, and invert DNA in vitro and in bacteria. Here, we report the discovery and engineering of IS622, a simple two-component system capable of universal DNA rearrangements of the human genome. We define strategies for the optimal application of bridge systems, leveraging mechanistic insights to improve their targeting specificity. Through rational engineering of the IS622 bridge RNA and deep mutational scanning of its recombinase, we achieve up to 20% insertion efficiency into the human genome and genome-wide specificity as high as 82%. We further demonstrate intra-chromosomal inversion and excision, mobilizing up to 0.93 megabases of DNA. Finally, we provide proof-of-concept for excision of a gene regulatory region or expanded repeats relevant for the treatment of genetic diseases.

bioengineering↗

bioRxiv2025

scBaseCamp: An AI agent-curated, uniformly processed, and continually expanding single cell data repository

Single-cell RNA sequencing has transformed cell biology by enabling precise transcriptomic measurements of individual cells. The Sequence Read Archive (SRA) is the largest public repository of sequencing reads, yet much of it remains underutilized due to unstandardized metadata and the cost of processing reads. Here, we introduce scBaseCount, a single-cell RNA sequencing database that leverages an AI agent to automate discovery and metadata extraction, and standardize data processing. Built by directly mining all 10x Genomics datasets from SRA, scBaseCount is the largest freely accessible public repository of single-cell gene expression data, comprising over 502 million cells across 27 organisms and 75 tissues, offering an unbiased view of the composition of data within SRA. Uniform processing enables measurement of both intronic and exonic reads, non-coding gene expression and improves alignment across experiments as well as the performance of AI models trained on this phenotypically diverse data. Moreover, scBaseCount provides a blueprint for how AI can be leveraged to curate and autonomously update large biological data repositories.

bioinformatics↗

bioRxiv2025

Genome modeling and design across all domains of life with Evo 2

All of life encodes information with DNA. While tools for sequencing, synthesis, and editing of genomic code have transformed biological research, intelligently composing new biological systems would also require a deep understanding of the immense complexity encoded by genomes. We introduce Evo 2, a biological foundation model trained on 9.3 trillion DNA base pairs from a highly curated genomic atlas spanning all domains of life. We train Evo 2 with 7B and 40B parameters to have an unprecedented 1 million token context window with single-nucleotide resolution. Evo 2 learns from DNA sequence alone to accurately predict the functional impacts of genetic variation--from noncoding pathogenic mutations to clinically significant BRCA1 variants--without task-specific finetuning. Applying mechanistic interpretability analyses, we reveal that Evo 2 autonomously learns a breadth of biological features, including exon-intron boundaries, transcription factor binding sites, protein structural elements, and prophage genomic regions. Beyond its predictive capabilities, Evo 2 generates mitochondrial, prokaryotic, and eukaryotic sequences at genome scale with greater naturalness and coherence than previous methods. Guiding Evo 2 via inference-time search enables controllable generation of epigenomic structure, for which we demonstrate the first inference-time scaling results in biology. We make Evo 2 fully open, including model parameters, training code, inference code, and the OpenGenome2 dataset, to accelerate the exploration and design of biological complexity.

genomics↗

bioRxiv2022

Self-cleaving 2A peptides allow for expression of multiple genes in Dictyostelium discoideum

The social amoeba Dictyostelium discoideum is a model for a wide range of biological processes including chemotaxis, cell-cell communication, phagocytosis, and development. Interrogating these processes with modern genetic tools often requires the expression of multiple transgenes. While it is possible to transfect multiple transcriptional units, the use of separate promoters and terminators for each gene leads to large plasmid sizes and possible interference between units. In other eukaryotic systems this challenge has been addressed through polycistronic expression mediated by 2A viral peptides, permitting efficient, co-regulated gene expression. Here, we screen the most commonly used 2A peptides, porcine teschovirus-1 2A (P2A), Thosea asigna virus 2A (T2A), equine rhinitis A virus 2A (E2A), and foot-and-mouth disease virus 2A (F2A), for activity in D. discoideum. We find that these 2A sequences are effective but that, unexpectedly, combining the coding sequences of two proteins into a single transcript leads to notable decreases in expression level with strain-dependent magnitudes, suggesting unexplored constraints of gene expression in D. discoideum that merit further investigation. Our results show that P2A is the optimal sequence for polycistronic expression, opening up new possibilities for genetic engineering in this model system.

molecular biology↗

bioRxiv2020

Decoupling transcription factor expression and activity enables dimmer-switch gene regulation

Our mechanistic understanding of how gene regulatory networks can achieve complex biological responses is limited. We find that the galactose-responsive pathway in S. cerevisiae has a complex behavior - the decision to induce is controlled independently from the induction level, resembling a mechanical dimmer-switch. Surprisingly, this behavior is not achieved directly through chromatin regulation or combinatorial control at galactose-responsive promoters. Instead, this behavior is achieved by hierarchical regulation of the expression and catalytic activity of a single transcription factor. This genetic motif allows evolution to independently act on both properties, providing a means to tune both resource allocation strategies in dynamic multi-input environments on both physiological and evolutionary time-scales. Hierarchical regulation is ubiquitous and thus dimmer-switch regulation is likely a key feature of many biological systems.

systems biology↗

bioRxiv2018

Dissecting the sharp response of a canonical developmental enhancer reveals multiple sources of cooperativity

Developmental enhancers integrate graded concentrations of input transcription factors (TFs) to create sharp gene expression boundaries. Here we examine the hunchback P2 (HbP2) enhancer which drives a sharp expression pattern in the Drosophila blastoderm embryo in response to the transcriptional activator Bicoid (Bcd). We systematically interrogate cis and transfactors that influence the shape and position of expression driven by HbP2, and find that the prevailing model, based on cooperative binding of Bcd to HbP2 is not adequate. We demonstrate that other proteins, such as pioneer factors, mediator and histone modifiers influence the shape and position of the HbP2 expression pattern. By comparing our results to theory, we assess how higher-order cooperativity and energy expenditure impact boundary location and sharpness. Our results emphasize that the bacterial view of transcription regulation, where pairwise interactions between regulatory proteins dominate, must be re-examined in animals, where multiple molecular mechanisms collaborate to shape the gene regulatory function.

genetics↗