Biology subjects

Winter, P. S.

Publications and source records attributed to Winter, P. S..

13 recordsLinked to original sources

bioRxiv2026

Single-cell perturbation prediction near experimental reproducibility with GENESIS

Single-cell perturbation atlases now measure cellular responses across millions of cells, but the number of possible combinations of biological contexts, perturbations, doses and other experimental variables make exhaustive measurements infeasible. Here, we present GENESIS (Generative ENErgy-based Single-cell Intervention Simulator), a generative framework that predicts single-cell response distributions directly from the covariates defining an experimental condition. GENESIS uses a lightweight neural network trained with the energy score, a strictly proper scoring rule for distributions. We introduce an experiment-calibrated evaluation framework in which prediction error is interpreted relative to the reproducibility of repeated measurements of the same condition. Across the Tahoe-100M small-molecule atlas and a large PBMC cytokine atlas, GENESIS significantly outperforms existing generative models and statistical baselines and approaches experimental reproducibility on independently repeated conditions. Importantly, GENESIS remains accurate when trained on only a fraction of the available experimental conditions or cells, and generated profiles recover experimentally observed context-dependent transcriptional programs. At comparable training-cell budgets, GENESIS achieves more accurate predictions when measurements span more experimental conditions rather than profiling fewer conditions more deeply. Together, these findings suggest a complementary strategy for perturbation studies: broader, lower-depth measurements can expand experimental coverage, while GENESIS augments sparsely sampled populations and predicts responses in unmeasured conditions.

bioinformatics↗

bioRxiv2026

Direct comparison of CRISPR knockout and interference with Perturb-seq

CRISPR knockout (CRISPRko) and CRISPR interference (CRISPRi) are two workhorse technologies for loss-of-function studies, yet direct comparisons between the two are scant relative to their widespread adoption. Here, we establish benchmarking libraries for Cas9-based CRISPRko and CRISPRi screens using Perturb-seq as the read-out. For both modalities, we observe consistent transcriptional signatures among cells with the same genes perturbed, strong evidence of on-target signal. We also examine tradeoffs between modalities: while CRISPRi guides demonstrate heightened rates of off-target activity, we also observe artifacts stemming from the cellular response to double-stranded breaks with the use of CRISPRko. The libraries and analyses presented here will be a useful benchmarking and de-risking resource for any group preparing for a large-scale Perturb-seq screen.

genomics↗

bioRxiv2026

Beyond alignment: synergistic integration is required for multimodal cell foundation models

The vision of a "virtual cell"--a computational model that simulates biological function across modalities and scales--has become a defining goal in computational biology. While powerful unimodal foundation models exist, the lack of large-scale paired data prohibits the joint training of multimodal approaches. This scarcity favors compositional foundation models (CFMs): architectures that fuse frozen unimodal experts via a learned interface. However, it remains unclear when this multimodal fusion adds task-relevant information beyond the strongest unimodal representation and when it merely aggregates redundant signal. Here, we introduce the Synergistic Information Score (SIS), a metric grounded in partial information decomposition (PID), that quantifies the information gain achievable only through cross-modal interactions. Extending theoretical results from self-supervised learning, we show that standard alignment-based fusion objectives on frozen encoders inherently collapse to detecting linear redundancies, limiting their ability to capture nonlinear synergistic states. This distinction is directly relevant for tasks aiming to link tissue morphology and gene expression. Benchmarking ten fusion methods on spatial transcriptomics datasets, we use SIS to demonstrate that tasks dominated by linear redundancies are sufficiently served by unimodal baselines, whereas complex niche definitions benefit from synergy-aware integration objectives that enable cross-modal interactions beyond linear alignment. Finally, we perform a scaling analysis which highlights that fine-tuning a dominant unimodal expert is the most sample-efficient path for standard tasks, suggesting that the benefits of multimodal frameworks only emerge when tasks depend on information distributed across modalities. Together, these results establish that building towards a virtual cell will require a fundamental shift from alignment objectives that emphasize shared structure to synergy-maximizing integration that preserves and exploits complementary cross-modal signal.

bioinformatics↗

bioRxiv2025

Adaptive resampling for improved machine learning in imbalanced single-cell datasets

While machine learning models trained on single-cell transcriptomics data have shown great promise in providing biological insights, existing tools struggle to effectively model underrepresented and out-of-distribution cellular features or states. We present a generalizable Adaptive Resampling (AR) approach that addresses these limitations and enhances single-cell representation learning by resampling data based on its learned latent structure in an online, adaptive manner concurrent with model training. Experiments on gene expression reconstruction, cell type classification, and perturbation response prediction tasks demonstrate that the proposed AR training approach leads to significantly improved downstream performance across datasets and metrics. Additionally, it enhances the quality of learned cellular embeddings compared to standard training methods. Our results suggest that AR may serve as a valuable technique for improving representation learning and predictive performance in single-cell transcriptomic models.

bioinformatics↗

bioRxiv2025

Consequences of training data composition for deep learning models in single-cell biology

Foundation models for single-cell transcriptomics have the potential to augment (or replace) purpose-built tools for a variety of common analyses, especially when data are sparse. Recent work with large language models has shown that training data composition greatly shapes performance; however, to date, single-cell foundation models have ignored this aspect, opting instead to train on the largest possible corpus. We systematically investigate the consequences of training dataset composition on the behavior of deep learning models of single-cell transcriptomics, focusing on human hematopoiesis as a tractable model system and including cells from adult and developing tissues, disease states, and perturbation atlases. We find that (1) these models generalize poorly to unseen cell types, (2) adding malignant cells to a healthy cell training corpus does not necessarily improve modeling of unseen malignant cells, and (3) including an embryonic stem cell differentiation atlas during training improves performance on out-of-distribution tasks. Our results emphasize the importance of diverse training data and suggest strategies to optimize future single-cell foundation models.

genomics↗

bioRxiv2024

Evaluating the role of pre-training dataset size and diversity on single-cell foundation model performance

The success of transformer-based foundation models on natural language and images has motivated their use in single-cell biology. Single-cell foundation models have been trained on increasingly larger transcriptomic datasets, scaling from initial studies with 1 million cells to newer atlases with over 100 million cells. This study investigates the role of pre-training dataset size and diversity on the performance of single-cell foundation models on both zero-shot and fine-tuned tasks. Using a large corpus of 22.2 million cells, we pre-train a total of 400 models, which we evaluate by conducting 6,400 experiments. Our results show that current methods tend to plateau in performance with pre-training datasets that are only a fraction of the size of current training corpora.

genomics↗

bioRxiv2024

Mutation and cell state compatibility is required and targetable in Ph+ acute lymphoblastic leukemia minimal residual disease

Efforts to cure BCR::ABL1 B cell acute lymphoblastic leukemia (Ph+ ALL) solely through inhibition of ABL1 kinase activity have thus far been insufficient despite the availability of tyrosine kinase inhibitors (TKIs) with broad activity against resistance mutants. The mechanisms that drive persistence within minimal residual disease (MRD) remain poorly understood and therefore untargeted. Utilizing 13 patient-derived xenograft (PDX) models and clinical trial specimens of Ph+ ALL, we examined how genetic and transcriptional features co-evolve to drive progression during prolonged TKI response. Our work reveals a landscape of cooperative mutational and transcriptional escape mechanisms that differ from those causing resistance to first generation TKIs. By analyzing MRD during remission, we show that the same resistance mutation can either increase or decrease cellular fitness depending on transcriptional state. We further demonstrate that directly targeting transcriptional state-associated vulnerabilities at MRD can overcome BCR::ABL1 independence, suggesting a new paradigm for rationally eradicating MRD prior to relapse. Finally, we illustrate how cell mass measurements of leukemia cells can be used to rapidly monitor dominant transcriptional features of Ph+ ALL to help rationally guide therapeutic selection from low-input samples. HIGHLIGHTSO_LIRelapse after remission on TKI can harbor mutations in ABL1, RAS, or neither C_LIO_LIMutations and development-like cell state dictate fitness in residual disease C_LIO_LICo-targeting cell state and ABL1 markedly reduces MRD C_LIO_LIBiophysical measurements provide an integrative, rapid measurement of cell state C_LI

cancer biology↗

bioRxiv2024

A knockoff calibration method to avoid over-clustering in single-cell RNA-sequencing

Standard single-cell RNA-sequencing (scRNA-seq) pipelines nearly always include unsupervised clustering as a key step in identifying biologically distinct cell types. A follow-up step in these pipelines is to test for differential expression between the identified clusters. When algorithms over-cluster, downstream analyses will produce inflated P -values resulting in increased false discoveries. In this work, we present callback (Calibrated Clustering via Knockoffs): a new method for protecting against over-clustering by controlling for the impact of reusing the same data twice when performing differential expression analysis, commonly known as "double-dipping". Importantly, our approach can be applied to a wide range of clustering algorithms. Using real and simulated data, we show that callback provides state-of-the-art clustering performance and can rapidly analyze large-scale scRNA-seq studies, even on a personal laptop.

bioinformatics↗

bioRxiv2024

Scalable nonparametric clustering with unified marker gene selection for single-cell RNA-seq data

Clustering is commonly used in single-cell RNA-sequencing (scRNA-seq) pipelines to characterize cellular heterogeneity. However, current methods face two main limitations. First, they require user-specified heuristics which add time and complexity to bioinformatic workflows; second, they rely on post-selective differential expression analyses to identify marker genes driving cluster differences, which has been shown to be subject to inflated false discovery rates. We address these challenges by introducing nonparametric clustering of single-cell populations (NCLUSION): an infinite mixture model that leverages Bayesian sparse priors to identify marker genes while simultaneously performing clustering on single-cell expression data. NCLUSION uses a scalable variational inference algorithm to perform these analyses on datasets with up to millions of cells. Through simulations and analyses of publicly available scRNA-seq studies, we demonstrate that NCLUSION (i) matches the performance of other state-of-the-art clustering techniques with significantly reduced runtime and (ii) provides statistically robust and biologically relevant transcriptomic signatures for each of the clusters it identifies. Overall, NCLUSION represents a reliable hypothesis-generating tool for understanding patterns of expression variation present in single-cell populations.

bioinformatics↗

bioRxiv2023

Compressed phenotypic screens for complex multicellular models and high-content assays

High-throughput phenotypic screens leveraging biochemical perturbations, high-content readouts, and complex multicellular models could advance therapeutic discovery yet remain constrained by limitations of scale. To address this, we establish a method for compressing screens by pooling perturbations followed by computational deconvolution. Conducting controlled benchmarks with a highly bioactive small molecule library and a high-content imaging readout, we demonstrate increased efficiency for compressed experimental designs compared to conventional approaches. To prove generalizability, we apply compressed screening to examine transcriptional responses of patient-derived pancreatic cancer organoids to a library of tumor-microenvironment (TME)-nominated recombinant protein ligands. Using single-cell RNA-seq as a readout, we uncover reproducible phenotypic shifts induced by ligands that correlate with clinical features in larger datasets and are distinct from reference signatures available in public databases. In sum, our approach enables phenotypic screens that interrogate complex multicellular models with rich phenotypic readouts to advance translatable drug discovery as well as basic biology.

systems biology↗

bioRxiv2023

MCB-613 exploits a collateral sensitivity in drug-resistant EGFR-mutant non-small cell lung cancer through covalent inhibition of KEAP1

Targeted therapies have revolutionized cancer chemotherapy. Unfortunately, most patients develop multifocal resistance to these drugs within a matter of months. Here, we used a high-throughput phenotypic small molecule screen to identify MCB-613 as a compound that selectively targets EGFR-mutant, EGFR inhibitor-resistant non-small cell lung cancer (NSCLC) cells harboring diverse resistance mechanisms. Subsequent proteomic and functional genomic screens involving MCB-613 identified its target in this context to be KEAP1, revealing that this gene is selectively essential in the setting of EGFR inhibitor resistance. In-depth molecular characterization demonstrated that (1) MCB-613 binds KEAP1 covalently; (2) a single molecule of MCB-613 is capable of bridging two KEAP1 monomers together; and, (3) this modification interferes with the degradation of canonical KEAP1 substrates such as NRF2. Surprisingly, NRF2 knockout sensitizes cells to MCB-613, suggesting that the drug functions through modulation of an alternative KEAP1 substrate. Together, these findings advance MCB-613 as a new tool for exploiting the selective essentiality of KEAP1 in drug-resistant, EGFR-mutant NSCLC cells.

cancer biology↗

bioRxiv2021

Deciphering the tumor-specific immunopeptidome in vivo with genetically engineered mouse models

Effective immunosurveillance of cancer requires the presentation of peptide antigens on major histocompatibility complex Class I (MHC-I). Recent developments in proteomics have improved the identification of peptides that are naturally presented by MHC-I, collectively known as the "immunopeptidome". Current approaches to profile tumor immunopeptidomes have been limited to in vitro investigation, which fails to capture the in vivo repertoire of MHC-I peptides, or bulk tumor lysates, which are obscured by the lack of tumor-specific MHC-I isolation. To overcome these limitations, we report here the engineering of a Cre recombinase-inducible affinity tag into the endogenous mouse MHC-I gene and targeting of this allele to the KrasLSL-G12D/+; p53fl/fl (KP) mouse model (KP; KbStrep). This novel approach has allowed us to isolate tumor-specific MHC-I peptides from autochthonous pancreatic ductal adenocarcinoma (PDAC) and lung adenocarcinoma (LUAD) in vivo. With this powerful analytical tool, we were able to profile the evolution of the LUAD immunopeptidome through tumor progression and show that in vivo MHC-I presentation is shaped by post-translational mechanisms. We also uncovered novel, putative LUAD tumor associated antigens (TAAs). Many peptides that were recurrently presented in vivo exhibited very low expression of the cognate mRNA, provoking reconsideration of antigen prediction pipelines that triage peptides according to transcript abundance. Beyond cancer, the KbStrep allele is compatible with a broad range of Cre-driver lines to explore antigen presentation in vivo in the pursuit of understanding basic immunology, infectious disease, and autoimmunity.

cancer biology↗

bioRxiv2020

Transcriptional subtype-specific microenvironmental crosstalk and tumor cell plasticity in metastatic pancreatic cancer

Bulk transcriptomic studies have defined classical and basal-like gene expression subtypes in pancreatic ductal adenocarcinoma (PDAC) that correlate with survival and response to chemotherapy; however, the underlying mechanisms that govern these subtypes and their heterogeneity remain elusive. Here, we performed single-cell RNA-sequencing of 23 metastatic PDAC needle biopsies and matched organoid models to understand how tumor cell-intrinsic features and extrinsic factors in the tumor microenvironment (TME) shape PDAC cancer cell phenotypes. We identify a novel cancer cell state that co-expresses basal-like and classical signatures, demonstrates upregulation of developmental and KRAS-driven gene expression programs, and represents a transitional intermediate between the basal-like and classical poles. Further, we observe structure to the metastatic TME supporting a model whereby reciprocal intercellular signaling shapes the local microenvironment and influences cancer cell transcriptional subtypes. In organoid culture, we find that transcriptional phenotypes are plastic and strongly skew toward the classical expression state, irrespective of genotype. Moreover, we show that patient-relevant transcriptional heterogeneity can be rescued by supplementing organoid media with factors found in the TME in a subtype-specific manner. Collectively, our study demonstrates that distinct microenvironmental signals are critical regulators of clinically relevant PDAC transcriptional states and their plasticity, identifies the necessity for considering the TME in cancer modeling efforts, and provides a generalizable approach for delineating the cell-intrinsic versus -extrinsic factors that govern tumor cell phenotypes.

cancer biology↗