Biology subjects

Hsu, P. D.

Publications and source records attributed to Hsu, P. D..

At least 19 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

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

Scalable probe-based single-cell transcriptional profiling for virtual cell perturbation mapping and synthetic biology phenotyping

Large-scale single-cell transcriptional phenotyping of genetic perturbations (perturb-seq) links genes to phenotypes and should enable virtual cell predictive modeling and cellular engineering. However, current perturb-seq single-cell methods are costly, information sparse and require barcodes for many applications. We developed ProPer-seq, a perturb-seq method that uses multiplexed custom DNA probe panels to measure and phenotype synthetic biology perturbations at single-cell resolution without barcodes, including multidomain proteins and sgRNAs. ProPer-seq faithfully reproduces gold-standard perturb-seq phenotypes while achieving 4-fold cost reduction and 50% increased gene detection per cell. As a scalable fixed-cell profiling method, ProPer-seq enables atlas-scale profiling for virtual-cell initiatives and demonstrates data quality suitable for training and validating predictive models. Lastly, ProPer-seqs targeted detection of modular transgenes enables library-on-library perturbation profiling of combinatorial synthetic protein design spaces. We applied this to 3,550 sgRNA x dCas9 effector combinations as well as 260 CAR x ORF combinations dynamically profiled in primary T cells, revealing principles of transcriptional control and cell state modulation by multidomain synthetic transgenes.

synthetic biology↗

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

Gene Editing for ATXN3 Inactivation in Machado-Joseph disease: CRISPR-Cas9 as a Therapeutic Alternative to TALEN-Induced Toxicity

Machado-Joseph disease (MJD) is an autosomal dominantly-inherited neurodegenerative disorder, caused by an over-repetition of the polyglutamine-codifying region in the ATXN3 gene. Strategies based on the suppression of the deleterious gene products have demonstrated promising results in pre-clinical studies. Nonetheless, these strategies do not target the root cause of the disease. In order to prevent the downstream toxic pathways, our goal was to develop gene editing-based strategies to permanently inactivate the human ATXN3 gene. TALENs and CRISPR-Cas9 systems were designed to target exon 2 of this gene and functional characterization was performed in a human cell line. After the demonstration of TALENs and CRISPR-Cas9 efficiency on gene disruption, a sequence of each system was selected for further in vivo experiments. Although both TALENs and CRISPR-Cas9 systems led to a drastic reduction of ATXN3 aggregates in the striatum of a lentiviral-based mouse model of MJD/SCA3, only CRISPR-Cas9 system allowed the improvement of key neuropathological markers of the disease. Importantly, the administration of the engineered system in YAC-MJD84.2/84.2 mice mediated a delay in disease progression, when compared with non-treated littermates. These data provide the first in vivo evidence of the efficacy of a CRISPR-Cas9-based approach to permanently inactivate the ATXN3 gene in the brain of two mouse models of the disease, supporting its potential as a new therapeutic avenue in the context of MJD/SCA3.

neuroscience↗

bioRxiv2024

Site-specific DNA insertion into the human genome with engineered recombinases

Technologies for precisely inserting large DNA sequences into the genome are critical for diverse research and therapeutic applications. Large serine recombinases (LSRs) can mediate direct, site-specific genomic integration of multi-kilobase DNA sequences without a pre-installed landing pad, but current approaches suffer from low insertion rates and high off-target activity. Here, we present a comprehensive engineering roadmap for the joint optimization of DNA recombination efficiency and specificity. We combined directed evolution, structural analysis, and computational models to rapidly identify additive mutational combinations. We further enhanced performance through donor DNA optimization and dCas9 fusions, enabling simultaneous target and donor recruitment. Top engineered LSR variants achieved up to 53% integration efficiency and 97% genome-wide specificity at an endogenous human locus, and effectively integrated large DNA cargoes (up to 12 kb tested) for stable expression in challenging cell types, including non-dividing cells, human embryonic stem cells, and primary human T cells. This blueprint for rational engineering of DNA recombinases enables precise genome engineering without the generation of double-stranded breaks.

bioengineering↗

bioRxiv2024

A combinatorial domain screening platform reveals epigenetic effector interactions for transcriptional perturbation

Epigenetic regulation involves the coordinated interplay of diverse proteins. To systematically explore these combinations, we present COMBINE (combinatorial interaction exploration), a high-throughput platform that tests over 50,000 pairs of epigenetic effector domains up to 2,094 amino acids in length for their ability to modulate endogenous human gene transcription. COMBINE revealed diverse synergistic and antagonistic interactions between epigenetic protein domains, including a potent KRAB-L3MBTL3 fusion that enhanced gene silencing up to 34-fold in dose-limited conditions and enabled robust bidirectional CRISPR perturbation. Inducible screening showed DNA methylation modifiers are essential for epigenetic memory, with distinct combinations driving long-term silencing, repression, or activation. Notably, we identified TET1-based combinations that induce hit-and-run upregulation for over 50 days, demonstrating long-term transcriptional activation. This systematic analysis of pairwise domain interactions provides a rich resource for understanding epigenetic crosstalk and developing next-generation epigenome editing tools. More broadly, COMBINE offers a generalizable platform to functionally characterize combinatorial biological processes at scale.

bioengineering↗

bioRxiv2024

Simulating 500 million years of evolution with a language model

More than three billion years of evolution have produced an image of biology encoded into the space of natural proteins. Here we show that language models trained on tokens generated by evolution can act as evolutionary simulators to generate functional proteins that are far away from known proteins. We present ESM3, a frontier multimodal generative language model that reasons over the sequence, structure, and function of proteins. ESM3 can follow complex prompts combining its modalities and is highly responsive to biological alignment. We have prompted ESM3 to generate fluorescent proteins with a chain of thought. Among the generations that we synthesized, we found a bright fluorescent protein at far distance (58% identity) from known fluorescent proteins. Similarly distant natural fluorescent proteins are separated by over five hundred million years of evolution.

synthetic biology↗

bioRxiv2024

Designing Cell-Type-Specific Promoter Sequences Using Conservative Model-Based Optimization

Gene therapies have the potential to treat disease by delivering therapeutic genetic cargo to disease-associated cells. One limitation to their widespread use is the lack of short regulatory sequences, or promoters, that differentially induce the expression of delivered genetic cargo in target cells, minimizing side effects in other cell types. Such cell-type-specific promoters are difficult to discover using existing methods, requiring either manual curation or access to large datasets of promoter-driven expression from both targeted and untargeted cells. Model-based optimization (MBO) has emerged as an effective method to design biological sequences in an automated manner, and has recently been used in promoter design methods. However, these methods have only been tested using large training datasets that are expensive to collect, and focus on designing promoters for markedly different cell types, overlooking the complexities associated with designing promoters for closely related cell types that share similar regulatory features. Therefore, we introduce a comprehensive framework for utilizing MBO to design promoters in a data-efficient manner, with an emphasis on discovering promoters for similar cell types. We use conservative objective models (COMs) for MBO and highlight practical considerations such as best practices for improving sequence diversity, getting estimates of model uncertainty, and choosing the optimal set of sequences for experimental validation. Using three relatively similar blood cancer cell lines (Jurkat, K562, and THP1), we show that our approach discovers many novel cell-type-specific promoters after experimentally validating the designed sequences. For K562 cells, in particular, we discover a promoter that has 75.85% higher cell-type-specificity than the best promoter from the initial dataset used to train our models.

synthetic biology↗

bioRxiv2024

Sequence modeling and design from molecular to genome scale with Evo

The genome is a sequence that completely encodes the DNA, RNA, and proteins that orchestrate the function of a whole organism. Advances in machine learning combined with massive datasets of whole genomes could enable a biological foundation model that accelerates the mechanistic understanding and generative design of complex molecular interactions. We report Evo, a genomic foundation model that enables prediction and generation tasks from the molecular to genome scale. Using an architecture based on advances in deep signal processing, we scale Evo to 7 billion parameters with a context length of 131 kilobases (kb) at single-nucleotide, byte resolution. Trained on 2.7M prokaryotic and phage genomes, Evo can generalize across the three fundamental modalities of the central dogma of molecular biology to perform zero-shot function prediction that is competitive with, or outperforms, leading domain-specific language models. Evo also excels at multi-element generation tasks, which we demonstrate by generating synthetic CRISPR-Cas molecular complexes and entire transposable systems for the first time. Using information learned over whole genomes, Evo can also predict gene essentiality at nucleotide resolution and can generate coding-rich sequences up to 650 kb in length, orders of magnitude longer than previous methods. Advances in multi-modal and multiscale learning with Evo provides a promising path toward improving our understanding and control of biology across multiple levels of complexity.

synthetic biology↗

bioRxiv2024

Rewriting endogenous human transcripts with trans-splicing

Splicing bridges the gap between static DNA sequence and the diverse and dynamic set of protein products that execute a genes biological functions. While exon skipping technologies enable influence over splice site selection, many desired perturbations to the transcriptome require replacement or addition of exogenous exons to target mRNAs: for example, to replace disease-causing exons, repair truncated proteins, or engineer protein fusions. Here, we report the development of RNA-guided trans-splicing with Cas editor (RESPLICE), inspired by the rare, natural process of trans-splicing that joins exons from two distinct primary transcripts. RESPLICE uses two orthogonal RNA-targeting CRISPR effectors to co-localize a trans-splicing pre-mRNA and to inhibit the cis-splicing reaction, respectively. We demonstrate efficient, specific, and programmable trans-splicing of multi-kilobase RNA cargo into nine endogenous transcripts across two human cell types, achieving up to 45% trans-splicing efficiency in bulk, or 90% when sorting for high effector expression. Our results present RESPLICE as a new mode of RNA editing for fine-tuned and transient control of cellular programs without permanent alterations to the genetic code.

molecular biology↗

bioRxiv2024

Bridge RNAs direct modular and programmable recombination of target and donor DNA

Genomic rearrangements, encompassing mutational changes in the genome such as insertions, deletions, or inversions, are essential for genetic diversity. These rearrangements are typically orchestrated by enzymes involved in fundamental DNA repair processes such as homologous recombination or in the transposition of foreign genetic material by viruses and mobile genetic elements (MGEs). We report that IS110 insertion sequences, a family of minimal and autonomous MGEs, express a structured non-coding RNA that binds specifically to their encoded recombinase. This bridge RNA contains two internal loops encoding nucleotide stretches that base-pair with the target DNA and donor DNA, which is the IS110 element itself. We demonstrate that the target-binding and donor-binding loops can be independently reprogrammed to direct sequence-specific recombination between two DNA molecules. This modularity enables DNA insertion into genomic target sites as well as programmable DNA excision and inversion. The IS110 bridge system expands the diversity of nucleic acid-guided systems beyond CRISPR and RNA interference, offering a unified mechanism for the three fundamental DNA rearrangements required for genome design.

genetics↗

bioRxiv2023

Pretraining strategies for effective promoter-driven gene expression prediction

The ability to deliver genetic cargo to human cells is enabling rapid progress in molecular medicine, but designing this cargo for precise expression in specific cell types is a major challenge. Expression is driven by regulatory DNA sequences within short synthetic promoters, but relatively few of these promoters are cell-type-specific. The ability to design cell-type-specific promoters using model-based optimization would be impactful for research and therapeutic applications. However, models of expression from short synthetic promoters (promoter-driven expression) are lacking for most cell types due to insufficient training data in those cell types. Although there are many large datasets of both endogenous expression and promoter-driven expression in other cell types, which provide information that could be used for transfer learning, transfer strategies remain largely unexplored for predicting promoter-driven expression. Here, we propose a variety of pretraining tasks, transfer strategies, and model architectures for modelling promoter-driven expression. To thoroughly evaluate various methods, we propose two benchmarks that reflect data-constrained and large dataset settings. In the data-constrained setting, we find that pretraining followed by transfer learning is highly effective, improving performance by 24 - 27%. In the large dataset setting, transfer learning leads to more modest gains, improving performance by up to 2%. We also propose the best architecture to model promoter-driven expression when training from scratch. The methods we identify are broadly applicable for modelling promoter-driven expression in understudied cell types, and our findings will guide the choice of models that are best suited to designing promoters for gene delivery applications using model-based optimization. Our code and data are available at https://github.com/anikethjr/promoter_models.

genomics↗

bioRxiv2022

Kinase-independent activity of DYRK1A promotes viral entry of highly pathogenic human coronaviruses

Identifying host genes essential for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has the potential to reveal novel drug targets and further our understanding of coronavirus disease 2019 (COVID-19). We previously performed a genome-wide CRISPR/Cas9 screen to identify pro-viral host factors for highly pathogenic human coronaviruses. Very few host factors were required by diverse coronaviruses across multiple cell types, but DYRK1A was one such exception. Although its role in coronavirus infection was completely unknown, DYRK1A encodes Dual Specificity Tyrosine Phosphorylation Regulated Kinase 1A and regulates cell proliferation, and neuronal development, among other cellular processes. Interestingly, individuals with Down syndrome overexpress DYRK1A 1.5-fold and exhibit 5-10x higher hospitalization and mortality rates from COVID-19 infection. Here, we demonstrate that DYRK1A regulates ACE2 and DPP4 transcription independent of its catalytic kinase function to support SARS-CoV, SARS-CoV-2, and MERS-CoV entry. We show that DYRK1A promotes DNA accessibility at the ACE2 promoter and a putative distal enhancer, facilitating transcription and gene expression. Finally, we validate that the pro-viral activity of DYRK1A is conserved across species using cells of monkey and human origin and an in vivo mouse model. In summary, we report that DYRK1A is a novel regulator of ACE2 and DPP4 expression that may dictate susceptibility to multiple highly pathogenic human coronaviruses. Whether DYRK1A overexpression contributes to heightened COVID-19 severity in individuals with Down syndrome through ACE2 regulation warrants further future investigation.

microbiology↗

bioRxiv2021

Large-scale discovery of recombinases for integrating DNA into the human genome

Recent microbial genome sequencing efforts have revealed a vast reservoir of mobile genetic elements containing integrases that could be useful genome engineering tools. Large serine recombinases (LSRs), such as Bxb1 and PhiC31, are bacteriophage-encoded integrases that can facilitate the insertion of phage DNA into bacterial genomes. However, only a few LSRs have been previously characterized and they have limited efficiency in human cells. Here, we developed a systematic computational discovery workflow that identifies thousands of new LSRs and their cognate DNA attachment sites by. We validate this approach via experimental characterization of LSRs in human cells, leading to three classes of LSRs distinguished from one another by their efficiency and specificity. We identify landing pad LSRs that efficiently integrate into synthetically installed attachment sites orthogonal to the human genome, human genome-targeting LSRs with computationally predictable pseudosites, and multi-targeting LSRs that can unidirectionally integrate cargos at with similar efficiency and superior specificity to commonly used transposases. LSRs from each category were functionally characterized in human cells, overall achieving up to 7-fold higher plasmid recombination than Bxb1 and genome insertion efficiencies of 40-70% with cargo sizes over 7 kb. Overall, we establish a paradigm for large-scale discovery of microbial recombinases and reconstruction of their target sites directly from microbial sequencing data. This strategy provides a rich resource of over 60 experimentally characterized LSRs that can function in human cells and thousands of additional candidates for large-payload genome editing without exposed DNA double-stranded breaks.

bioengineering↗

bioRxiv2021

Deep learning of Cas13 guide activity from high-throughput gene essentiality screening

Transcriptome engineering technologies that can effectively and precisely perturb mammalian RNAs are needed to accelerate biological discovery and RNA therapeutics. However, the broad utility of programmable CRISPR-Cas13 ribonucleases has been hampered by an incomplete understanding of the design rules governing guide RNA activity as well as cellular toxicity resulting from off-target or collateral RNA cleavage. Here, we sought to characterize and develop Cas13d systems for efficient and specific RNA knockdown with low cellular toxicity in human cells. We first quantified the performance of over 127,000 RfxCas13d (CasRx) guide RNAs in the largest-scale screen to date and systematically evaluated three linear, two ensemble, and two deep learning models to build a guide efficiency prediction algorithm validated across multiple human cell types in orthogonal validation experiments (https://www.RNAtargeting.org). Deep learning model interpretation revealed specific sequence motifs at spacer position 15-24 along with favored secondary features for highly efficient guides. We next identified 46 novel Cas13d orthologs through metagenomic mining for activity and cytotoxicity screening, discovering that the metagenome-derived DjCas13d ortholog achieves low cellular toxicity and high transcriptome-wide specificity when deployed against high abundance transcripts or in sensitive cell types, including human embryonic stem cells, neural progenitor cells, and neurons. Finally, our Cas13d guide efficiency model successfully generalized to DjCas13d, highlighting the utility of a comprehensive approach combining machine learning with ortholog discovery to advance RNA targeting in human cells.

bioengineering↗

bioRxiv2021

Genome-wide, bidirectional CRISPR screens identify mucins as critical host factors modulating SARS-CoV-2 infection

SARS-CoV-2 can cause a range of symptoms in infected individuals, from mild respiratory illness to acute respiratory distress syndrome. A systematic understanding of the host factors mediating viral infection or restriction is critical to elucidate SARS-CoV-2 host-pathogen interactions and the progression of COVID-19. To this end, we conducted genome-wide CRISPR knockout and activation screens in human lung epithelial cells with endogenous expression of the SARS-CoV-2 entry factors ACE2 and TMPRSS2. These screens uncovered proviral and antiviral host factors across highly interconnected host pathways, including components implicated in clathrin transport, inflammatory signaling, cell cycle regulation, and transcriptional and epigenetic regulation. We further identified mucins, a family of high-molecular weight glycoproteins, as a prominent viral restriction network. We demonstrate that multiple membrane-anchored mucins are critical inhibitors of SARS-CoV-2 entry and are upregulated in response to viral infection. This functional landscape of SARS-CoV-2 host factors provides a physiologically relevant starting point for new host-directed therapeutics and suggests interactions between SARS-CoV-2 and airway mucins of COVID-19 patients as a host defense mechanism.

genetics↗