Biology subjects

Getz, G.

Publications and source records attributed to Getz, G..

At least 19 recordsLinked to original sources

bioRxiv2026

Clustered structural variant hotspots enable oncogenic addiction and plasticity in osteosarcoma

The genomic landscape of osteosarcoma, the most common bone cancer worldwide, is among the most structurally complex of all human malignancies. The identification of recurrent and functionally consequential patterns has thus remained a challenge. Across 236 whole-genome sequencing osteosarcoma samples, we uncovered five genomic hotspots of clustered structural variation collectively altered in 58% of tumors. Four were associated with amplification of oncogenes ( MYC , CCND3 , CCNE1 , CDK4) , while the fifth mapped largely upstream of TP53 . Hotspot events showed coordinated patterns of co-occurrence and mutual exclusivity with each other and with tumor suppressor alterations, suggesting genomic context-specific selection. We found localized transcriptional dysregulation at hotspot event loci, and single cells harboring these events converged on a neural crest-like program, linking these structural alterations to a less differentiated cell state. These events were also detectable non-invasively through liquid biopsies and displayed ongoing structural evolution throughout disease progression, a finding with potential clinical utility. Our results provide novel insight into how complex rearrangements shape oncogenesis in osteosarcoma, with broader relevance to other cancers characterized by complex genomes.

cancer biology↗

bioRxiv2026

Comprehensive characterization of genomic, transcriptomic and epigenomic artifacts introduced in formalin-fixed, paraffin-embedded tissues.

Genomic, transcriptomic and epigenomic characterization has accelerated the discovery of clinically-relevant alterations in cancer, predominantly using fresh frozen (FF) specimens. However, clinical molecular pathology laboratories prefer formalin-fixed paraffin-embedded (FFPE) methods, known to introduce artifacts at the nucleic acid level, over fresh frozen methods. Extending the multi-platform analysis to FFPE specimens for comprehensive clinical molecular diagnosis requires a thorough understanding of the consequence of formalin-fixation. We present a detailed multi-platform characterization of FFPE preservation using paired FF specimens as the 'gold standard'. DNA and RNA were obtained from 38 patients across 6 cancer types using a FFPE optimized co-isolation. The impact of FFPE on exome sequencing was dependent on filtering, where a minimum coverage or supporting read filter can mitigate FFPE-specific false positives. Copy number alterations, MSI assessment, mutational signatures, and DNA methylation were comparable between FFPE and FF. FFPE biases in RNA expression can be overcome when using biology-relevant genes and we describe a novel consequence of FFPE on miRNA species diversity. Collectively, this data provides a broad view of FFPE artifact and offers best practices for overcome these biases.

bioinformatics↗

bioRxiv2026

Tabular foundation model predicts alternative lengthening of telomeres (ALT) and identifies SMARCAL1 as a target in ALT-driven cancers

Alternative lengthening of telomeres (ALT) is a telomerase-independent pathway used by aggressive cancers to maintain their replicative immortality. Because ALT is absent from normal human cells, it is an appealing target for cancer therapy, but the lack of ability to determine ALT status at scale has hindered discovery. Here, we developed ALTitude, a machine learning method from a tabular foundation model that infers ALT from cell line whole genome sequencing data, without need for paired germline analysis. We deployed ALTitude across the DepMap, doubling the number of known ALT+ cancer models. Systematic integration of ALTitude with CRISPR-Cas9 screens yielded the selective dependency on SMARCAL1 in ALT+ cell lines, where we show it stabilizes the ALT phenotype. Acute depletion of SMARCAL1 leads to G2/M arrest, mitotic catastrophe, and cell death. These data provide a valuable resource for studying ALT-related genomic features and present SMARCAL1 as a therapeutic target for ALT+ malignancies.

cancer biology↗

bioRxiv2025

A Diffusion-Based Autoencoder for Learning Patient-Level Representations from Single-Cell Data

Single-cell RNA sequencing (scRNA-seq) offers insights into cellular heterogeneity and tissue composition, yet leveraging this data for patient-level clinical predictions remains challenging due to the set-structured nature of single-cell data, as well as the scarcity of labeled samples. To address these challenges, we introduce scSet, a diffusion-based autoencoder that learns patient-level representations from sets of single-cell transcriptomes. Our method uses a transformer-based encoder to process variably sized and unordered cell inputs, coupled with a conditional diffusion decoder for self-supervised learning on unlabeled data. By pre-training on large-scale unlabeled datasets, scSet generates robust patient representations that can be fine-tuned for downstream clinical prediction tasks. We demonstrate the effectiveness of scSet patient embeddings for clinical prediction across multiple real-world datasets, where they outperform existing patient representations, even with limited labeled data. This work represents an important step toward bridging the gap between single-cell resolution and patient-level insights. Code is available at https://github.com/clinicalml/scset.

bioinformatics↗

bioRxiv2025

SVelfie: A method for discovering cancer drivers based on enrichment of likely functional structural variants

Structural variants (SVs) can drive tumorigenesis, yet discovering SV cancer drivers remains challenging1. Here, we present SVelfie (Structural Variants enriched with likely functional/impactful events), a statistical method to infer driver genes from SVs detected across a cohort of cancer genomes, based on enrichment of likely functional events. When testing SVelfie on lymphoma samples from the Pan-Cancer Analysis of Whole Genomes (PCAWG)2, it corroborates known tumor suppressor genes and also yields novel driver candidates.

bioinformatics↗

bioRxiv2025

Single Cell RNA-sequencing of BCG naive and recurrent non-muscle invasive bladder cancer reveals a CD6/ALCAM mediated immune-suppressive pathway

Non-muscle invasive bladder cancer (NMIBC) represents 70-80% patients with newly diagnosed bladder cancer, and Bacillus Calmette-Guerin (BCG) remains a cornerstone treatment for intermediate-and high-risk NMIBC to prevent disease recurrence and progression. However, many patients experience recurrence after induction BCG, posing significant challenges in the management of the disease. We conducted single cell RNA sequencing on freshly collected NMIBC samples, distinguishing between those naive to BCG treatment and those that recurred post-BCG treatment. We observed a clear activation of inflammatory pathways across cell types during recurrence, but these were not associated with canonical immune checkpoint or T cell exhaustion phenotypes. Analysis of cell-to-cell communication revealed enhanced interactions between T cells and urothelial cells in BCG-recurrence, predominantly modulated by CD6 and ALCAM. Furthermore, we found CD6hi T cells to be immunosuppressed and enriched in recurrent samples, suggesting a potential role for CD6 as an immune evasion signal in NMIBC. SIGNIFICANCEThese findings uncover a novel mechanism responsible for bladder cancer recurrence after BCG treatment. Enhanced T cell-urothelial cell communication in recurrent tumors mediated by CD6 and ALCAM leads to an immunosuppressed state. Thus, CD6 may have potential as a therapeutic target to augment BCG response in non-muscle invasive bladder cancer.

cancer biology↗

bioRxiv2024

A unique interplay of access and selection shapes peritoneal metastasis evolution in colorectal cancer

Whether metastasis in humans can be accomplished by most primary tumor cells or requires the evolution of a specialized trait remains an open question. To evaluate whether metastases are founded by non-random subsets of primary tumor lineages requires extensive, difficult-to-implement sampling. We have realized an unusually dense multi-region sampling scheme in a cohort of 26 colorectal cancer patients with peritoneal metastases, reconstructing the evolutionary history of on average 28.8 tissue samples per patient with a microsatellite-based fingerprinting assay. To assess metastatic randomness, we evaluate inter- and intra-metastatic heterogeneity relative to the primary tumor and find that peritoneal metastases are more heterogeneous than liver metastases but less diverse than locoregional metastases. Metachronous peritoneal metastases exposed to systemic chemotherapy show significantly higher inter-lesion diversity than synchronous, untreated metastases. Projection of peritoneal metastasis origins onto a spatial map of the primary tumor reveals that they often originate at the deep-invading edge, in contrast to liver and lymph node metastases which exhibit no such preference. Furthermore, peritoneal metastases typically do not share a common subclonal origin with distant metastases in more remote organs. Synthesizing these insights into an evolutionary portrait of peritoneal metastases, we conclude that the peritoneal-metastatic process imposes milder selective pressures onto disseminating cancer cells than the liver-metastatic process. Peritoneal metastases unique evolutionary features have potential implications for staging and treatment.

genetics↗

bioRxiv2024

Generation of a biliary tract cancer cell line atlas reveals molecular subtypes and therapeutic targets

Biliary tract cancers (BTCs) are a group of deadly malignancies encompassing intrahepatic and extrahepatic cholangiocarcinoma, gallbladder carcinoma, and ampullary carcinoma. Here, we present the integrative analysis of 63 BTC cell lines via multi-omics clustering and genome- scale CRISPR screens, providing a platform to illuminate BTC biology and inform therapeutic development. We identify dependencies broadly enriched in BTC compared to other cancers as well as dependencies selective to the anatomic subtypes. Notably, cholangiocarcinoma cell lines are stratified into distinct lineage subtypes based on biliary or dual biliary/hepatocyte marker signatures, associated with dependency on specific lineage survival factors. Transcriptional analysis of patient specimens demonstrates the prognostic significance of these lineage subtypes. Additionally, we delineate strategies to enhance targeted therapies or to overcome resistance in cell lines with key driver gene mutations. Furthermore, clustering based on dependencies and proteomics data elucidates unexpected functional relationships, including a BTC subgroup with partial squamous differentiation. Thus, this cell line atlas reveals potential therapeutic targets in molecularly defined BTCs, unveils biologically distinct disease subtypes, and offers a vital resource for BTC research.

cancer biology↗

bioRxiv2023

scDiffEq: drift-diffusion modeling of single-cell dynamics with neural stochastic differential equations

Single-cell sequencing measurements facilitate the reconstruction of dynamic biology by capturing snapshot molecular profiles of individual cells. Cell fate decisions in development and disease are orchestrated through an intricate balance of deterministic and stochastic regulatory events. Drift-diffusion equations are effective in modeling single-cell dynamics from high-dimensional single-cell measurements. While existing solutions describe the deterministic dynamics associated with the drift term of these equations at the level of cell state, diffusion is modeled as a constant across cell states. To fully understand the dynamic regulatory logic in development and disease, models explicitly attuned to the balance between deterministic and stochastic biology are required. To address these limitations, we introduce scDiffEq, a generative framework for learning neural stochastic differential equations that approximate biologys deterministic and stochastic dynamics. Using lineage-traced single-cell data, we demonstrate that scDiffEq offers an improved reconstruction of cell trajectories and prediction of cell fate from multipotent progenitors during hematopoiesis. By imparting in silico perturbations to multipotent progenitor cells, we find that scDiffEq accurately recapitulates the dynamics of CRISPR-perturbed hematopoiesis. We generalize this approach beyond lineage-traced or multi-time point datasets to model the dynamics of single-cell data from a single time point. Using scDiffEq, we simulate high-resolution developmental cell trajectories, which can model their drift and diffusion, enabling us to study their time-dependent gene-level dynamics.

bioinformatics↗

bioRxiv2023

A Deep Dive into Single-Cell RNA Sequencing Foundation Models

Large-scale foundation models, which are pre-trained on massive, unlabeled datasets and subsequently fine-tuned on specific tasks, have recently achieved unparalleled success on a wide array of applications, including in healthcare and biology. In this paper, we explore two foundation models recently developed for single-cell RNA sequencing data, scBERT and scGPT. Focusing on the fine-tuning task of cell type annotation, we explore the relative performance of pre-trained models compared to a simple baseline, L1-regularized logistic regression, including in the few-shot setting. We perform ablation studies to understand whether pretraining improves model performance and to better understand the difficulty of the pre-training task in scBERT. Finally, using scBERT as an example, we demonstrate the potential sensitivity of fine-tuning to hyperparameter settings and parameter initializations. Taken together, our results highlight the importance of rigorously testing foundation models against well established baselines, establishing challenging fine-tuning tasks on which to benchmark foundation models, and performing deep introspection into the embeddings learned by the model in order to more effectively harness these models to transform single-cell data analysis. Code is available at https://github.com/clinicalml/sc-foundation-eval.

genomics↗

bioRxiv2023

Systematic characterization of therapeutic vulnerabilities in Multiple Myeloma with Amp1q reveals increased sensitivity to the combination of MCL1 and PI3K inhibitors

The development of targeted therapy for patients with Multiple Myeloma (MM) is hampered by the low frequency of actionable genetic abnormalities. Gain or amplification of chr1q (Amp1q) is the most frequent arm-level copy number gain in patients with MM, and it is associated with higher risk of progression and death despite recent advances in therapeutics. Thus, developing targeted therapy for patients with MM and Amp1q stands to benefit a large portion of patients in need of more effective management. Here, we employed large-scale dependency screens and drug screens to systematically characterize the therapeutic vulnerabilities of MM with Amp1q and showed increased sensitivity to the combination of MCL1 and PI3K inhibitors. Using single-cell RNA sequencing, we compared subclones with and without Amp1q within the same patient tumors and showed that Amp1q is associated with higher levels of MCL1 and the PI3K pathway. Furthermore, by isolating isogenic clones with different copy number for part of the chr1q arm, we showed increased sensitivity to MCL1 and PI3K inhibitors with arm-level gain. Lastly, we demonstrated synergy between MCL1 and PI3K inhibitors and dissected their mechanism of action in MM with Amp1q.

cancer biology↗

bioRxiv2023

Novel mechanism of MYC deregulation in Multiple Myeloma

MYC deregulation occurs in 67% of multiple myeloma (MM) cases and associates with progression and worse prognosis in MM. Enhanced MYC expression is known to be driven by translocation or amplification events, but it only occurs in 40% of MM patients. Here, we describe a new mechanism of MYC regulation, whereby epigenetic regulation of MYC by increased accessibility of a cell-type specific enhancer leads to increased MYC expression. We found enhancer activity does not associate with enhancer hijacking events. We identified specific binding of c-MAF, IRF4, and SPIB transcription factors to the enhancer can activate MYC. In addition, we discovered focal amplification of this specific enhancer in approximately 4% of MM patients. Together, our findings define a new epigenetic mechanism of MYC deregulation in MM beyond known translocations or amplifications and point to the importance of non-coding regulatory elements and their associated transcription factor networks as drivers of MM progression.

cancer biology↗

bioRxiv2022

Latent human herpesvirus 6 is reactivated in chimeric antigen receptor T cells

Cell therapies have yielded durable clinical benefits for patients with cancer, but the risks associated with the development of therapies from manipulated human cells are still being understood. For example, we currently lack a comprehensive understanding of the mechanisms of neurotoxicity observed in patients receiving T cell therapies, including recent reports of encephalitis caused by human herpesvirus 6 (HHV-6) reactivation1. Here, via petabase-scale viral RNA data mining, we examine the landscape of human latent viral reactivation and demonstrate that HHV-6B can become reactivated in human CD4+ T cells in standard in vitro cultures. Using single-cell sequencing, we identify a rare population of HHV-6 super-expressors (~1 in 300-10,000 cells) that possess high viral transcriptional activity in chimeric antigen receptor (CAR) T cell culture before spreading to infect other cells in vitro. Through the analysis of single-cell sequencing data from patients receiving cell therapy products that are FDA-approved2 or used in clinical studies3,4, we identify the presence of CAR+, HHV-6 super-expressor T cells in vivo. Together, our study implicates cell therapy products as a source of lytic HHV-6 reported in clinical trials1,5-7 and has broad implications for the design, production, and monitoring of cell therapies.

immunology↗

bioRxiv2022

Integrative Analysis of Checkpoint Blockade Response in Advanced Non-Small Cell Lung Cancer

Anti-PD-1/PD-L1 agents have transformed the treatment landscape of advanced non-small cell lung cancer (NSCLC). While our understanding of the biology underlying immune checkpoint blockade in NSCLC is still incomplete, studies to date have established predictive roles for PD-L1 tumor expression and tumor mutational burden (TMB). To expand our understanding of the molecular features underlying response to checkpoint inhibitors in NSCLC, we describe here the first joint analysis of the Stand Up 2 Cancer - Mark Foundation (SU2C-MARK) Cohort, a resource of whole exome and/or RNA sequencing from 393 patients with NSCLC treated with anti-PD-(L)1 therapy, along with matched clinical response annotation. We identify a number of associations between molecular features and outcome, including: 1) favorable (e.g., ATM altered), and unfavorable (e.g., TERT amplified) genomic subgroups, 2) distinct immune infiltration signatures associated with wound healing (unfavorable) and immune activation (favorable), and 3) a novel de-differentiated tumor-intrinsic subtype characterized by expression of endodermal lineage genes, immune activation, and enhanced response rate. Taken together, results from this cohort extend our understanding of NSCLC-specific predictors, providing a rich set of molecular and immunologic hypotheses with which to further our understanding of the biology of checkpoint blockade in NSCLC.

genomics↗

bioRxiv2022

High-resolution lung adenocarcinoma expression subtypes identify tumors with dependencies on MET, CDK4, CDK6, and PD-L1

Lung adenocarcinoma is one of the most common cancer types with various treatment modalities. However, better biomarkers to predict therapeutic response are still needed to improve precision medicine. We utilized a consensus hierarchical clustering approach on 509 LUAD cases from TCGA to identify five robust LUAD expression subtypes. We then integrated genomic (patient and cell line) and proteomic data to help define biomarkers of response to targeted therapies and immunotherapies. This approach defined subtypes with unique proteogenomic and dependency profiles. S4-associated cell lines exhibited specific vulnerability to CDK6 and CDK6-cyclin D3 complex gene, CCND3. S3 was characterized by dependency on CDK4, immune-related expression patterns, and altered MET signaling; experimental validation showed that S3-associated cell lines responded to MET inhibitors, leading to increased PD-L1 expression. We further identified genomic features in S3 and S4 as biomarkers for enabling clinical diagnosis of these subtypes. Overall, our consensus hierarchical clustering approach identified robust tumor expression subtypes, and our subsequent integrative analysis of genomics, proteomics, and CRISPR screening data revealed subtype-specific biology and vulnerabilities. Our lung adenocarcinoma expression subtypes and their biomarkers could help identify patients likely to respond to CDK4/6, MET, or PD-L1 inhibitors, potentially improving patient outcome. SignificanceThrough integrative analysis of genomic, proteomic, and drug dependency data, we identified robust lung adenocarcinoma expression subtypes and found subtype-specific biomarkers of response, including CDK4/6, MET, and PD-L1 inhibitors.

cancer biology↗

bioRxiv2021

Genetic Subtypes of Smoldering Multiple Myeloma are associated with Distinct Pathogenic Phenotypes and Clinical Outcomes

Smoldering multiple myeloma (SMM) is a precursor condition of multiple myeloma (MM) with significant heterogeneity in disease progression. Existing clinical models of progression risk do not fully capture this heterogeneity. Here we integrated 42 genetic alterations from 214 SMM patients using unsupervised binary matrix factorization (BMF) clustering and identified six distinct genetic subtypes. These subtypes were differentially associated with established MM-related RNA signatures, oncogenic and immune transcriptional profiles, and evolving clinical biomarkers. Three subtypes were associated with increased risk of progression to active MM in both the primary and validation cohorts, indicating they can be used to better predict high and low-risk patients within the currently used clinical risk stratification model.

cancer biology↗

bioRxiv2021

Single-nucleus cross-tissue molecular reference maps to decipher disease gene function

Understanding the function of genes and their regulation in tissue homeostasis and disease requires knowing the cellular context in which genes are expressed in tissues across the body. Single cell genomics allows the generation of detailed cellular atlases in human tissues, but most efforts are focused on single tissue types. Here, we establish a framework for profiling multiple tissues across the human body at single-cell resolution using single nucleus RNA-Seq (snRNA-seq), and apply it to 8 diverse, archived, frozen tissue types (three donors per tissue). We apply four snRNA-seq methods to each of 25 samples from 16 donors, generating a cross-tissue atlas of 209,126 nuclei profiles, and benchmark them vs. scRNA-seq of comparable fresh tissues. We use a conditional variational autoencoder (cVAE) to integrate an atlas across tissues, donors, and laboratory methods. We highlight shared and tissue-specific features of tissue-resident immune cells, identifying tissue-restricted and non-restricted resident myeloid populations. These include a cross-tissue conserved dichotomy between LYVE1- and HLA class II-expressing macrophages, and the broad presence of LAM-like macrophages across healthy tissues that is also observed in disease. For rare, monogenic muscle diseases, we identify cell types that likely underlie the neuromuscular, metabolic, and immune components of these diseases, and biological processes involved in their pathology. For common complex diseases and traits analyzed by GWAS, we identify the cell types and gene modules that potentially underlie disease mechanisms. The experimental and analytical frameworks we describe will enable the generation of large-scale studies of how cellular and molecular processes vary across individuals and populations.

cell biology↗

bioRxiv2021

Combined tumor and immune signals from genomes or transcriptomes predict outcomes of checkpoint inhibition in melanoma

Cancer immunotherapy with checkpoint blockade (CPB) leads to improved outcomes in melanoma and other tumor types, but a majority of patients do not respond. High tumor mutation burden (TMB) and high levels of tumor-infiltrating T cells have been associated with response to immunotherapy, but integrative models to predict clinical benefit using DNA or RNA alone have not been comprehensively explored. We sequenced DNA and RNA from melanoma patients receiving CPB, and aggregated previously published data, yielding whole exome sequencing data for 189 patients and bulk RNA sequencing data for 178 patients. Using these datasets, we derived genomic and transcriptomic factors that predict overall survival (OS) and response to immunotherapy. Using whole-exome DNA data alone, we calculated T cell burden (TCB) and B cell burden (BCB) based on rearranged TCR/Ig DNA sequences and found that patients whose melanomas have high TMB together with either high TCB or high BCB survived longer and had higher response rates as compared to patients with either low TMB or TCB/BCB. Next, using bulk RNA-Seq data, differential expression analysis identified 83 genes associated with high or low OS. By combining pairs of immune-expressed genes with tumor-expressed genes, we identified three gene pairs associated with response and survival (Bonferroni P<0.05). All three gene pair models were validated in an independent cohort (n=180) (Bonferroni P<0.05). The best performing gene pair model included the lymphocyte-expressed MAP4K1 (Mitogen- Activated Protein Kinase Kinase Kinase Kinase 1) combined with the transcription factor TBX3 (T-Box Transcription Factor 3) which is overexpressed in poorly differentiated melanomas. We conclude that RNA-based (MAP4K1&TBX3) or DNA-based (TCB&TMB) models combining immune and tumor measures improve predictions of outcome after checkpoint blockade in melanoma.

genomics↗