The poster session will take place on Monday, October 20th from 4:00 - 5:30 PM in the ARC Lobby. The session will be split so that odd-numbered posters will present for the first 45 minutes, and even-numbered posters will present for the second 45 minutes, to allow presenters an opportunity to enjoy the poster session as well.
Poster board size is 4'H x 6''W.
All Trainee (Undergrad, Grad Student and Postdoc) Poster presenters will be considered for a poster prize. Each poster will be evaluated by a PI and a Peer Junior Investigator Evaluator. Please review the poster judging criteria here.
1) Tracking Transcriptomic Drivers of Heterogeneous Resistance to Targeted Therapy in TNBC With Cell Barcodes
Michael Cotner, University of Texas at Austin
Triple negative breast cancer (TBNC) is marked by fewer standard-of-care treatment options and poorer treatment outcomes than other breast cancer subtypes, with approximately 40% of TNBC patients developing treatment resistance. High intratumoral heterogeneity, a characteristic of TNBC, leads to its difficulty in treatment and rapid acquisition of resistance. To investigate how this heterogeneity influences treatment response and resistance in TNBC, we employ ClonMapper, our DNA barcoding technology that utilizes integrated and heritable unique DNA barcodes, to track clonal cell populations across treatment. ClonMapper barcodes are identifiable in scRNA-seq, which enables tracking of clonal subpopulations and their transcriptomic diversity before and after treatment. Using a novel hierarchical Bayesian model conditioned on ClonMapper-enabled sequencing data, we quantify each barcoded clonal subpopulation’s diverse level of persistence during treatment, along with each subpopulation’s unique transcriptomic state before and after treatment. A predictive linear model fits treatment persistence as a function of pre-existing transcriptomic state and acquired transcriptomic changes, enabling the identification of genes whose heterogeneous expression across a tumor give rise to heterogeneous responses to treatment. We use our ClonMapper-guided model to investigate the treatment of TNBC tumor cell populations with three clinically-relevant targeted inhibitor chemotherapies, revealing how diversity in response to treatment within tumor cell populations arises from both the heterogenous transcriptomic landscape prior to treatment and the diverse transcriptomic trajectories taken by different cell subpopulations.
Other Authors: Amy Brock
2) Ras-dependent activation of BMAL2 regulates hypoxic metabolism in pancreatic cancer
Alvaro Curiel Garcia, Columbia University
To uncover novel drivers of pancreatic ductal adenocarcinoma (PDAC), we used regulatory network analysis to infer protein activity from gene expression data. Applying this to 200 human PDAC samples and 45 precursors, we identified master regulators linked to tumor initiation, progression, post-resection survival, and KRAS activity. BMAL2 emerged as the top regulator across all phenotypes. Though classically associated with circadian rhythm, BMAL2 showed strong links to hypoxia, a key PDAC hallmark. We confirmed its conservation among hypoxia-related genes and its suppression by RAF/MEK/ERK inhibitors. BMAL2 knockout in PDAC cells impaired viability, invasion, and glycolysis under hypoxia, and strikingly abolished HIF1A stabilization while increasing HIF2A. Metabolomic data showed reduced glycolytic activity in BMAL2-deficient cells. Preliminary xenograft studies showed diminished tumor growth with BMAL2 loss. These results position BMAL2 as a key RAS-regulated driver of hypoxic adaptation in PDAC and offer insight into distinct HIF1A/HIF2A regulation.
Other Authors: H. Carlo Maurer, Sam R. Holmstrom, Cristina Castillo, Carmine F. Palermo, Steven A. Sastra, Anthony Andren, Li Zhang, Tessa Y.S. Le Large, Irina Sagalovskiy, Winston Wong, Kaitlin Shaw, Jeanine Genkinger, Gulam A. Manji, Alina C. Iuga, Roland M. Schmid, Kristen Johnson, Michael A. Badgley, Costas A. Lyssiotis, Yatrik Shah, Andrea Califano, Kenneth P. Olive
3) scMINER2 - A causal feature learning framework to untangle cell-type-resolved hidden gene regulatory network (hGRN)
Yogesh Dhungana, St. Jude Children's Research Hospital
Single-cell technologies provide unprecedented resolution into immune cell biology, but they also intensify the curse of dimensionality, where sparsity and high dimensionality undermine traditional similarity and clustering approaches. Addressing this limitation requires not only dimensionality reduction but also mechanistic frameworks that preserve causal biological signals. To overcome these challenges, we developed scMINER2, a systems biology framework that simultaneously reconstructs hidden gene regulatory networks (hGRNs) and lineage trajectories from scRNA-seq and scATAC-seq data. By integrating transcriptional and chromatin accessibility profiles, scMINER2 identifies differentiation pathways and regulatory bottlenecks. Benchmarking against existing tools demonstrated that scMINER2 not only outperforms in accuracy but also enables deeper mechanistic interpretation. Applied to CAR-T cells, scMINER2 revealed a previously unrecognized subset of type-2 memory (T2M) cells. In contrast to historically suppressive type-2 programs, T2M cells emerged as long-lived, self-renewing, and indispensable drivers of CAR-T persistence, reframing type-2 immunity as a hidden dimension of therapeutic success in cancer immunotherapy. To validate this finding, we developed scPhenoMAP, a single-cell 2D embedding algorithm that projects T cells into a unified effector-state map across datasets. Using CAR-T scRNA-seq from global multi-center clinical trials spanning multiple continents, tumor types, and CAR constructs, scPhenoMAP consistently identified T2M enrichment in responders. This global reproducibility, further confirmed by preclinical experimental validation, underscores the universality and therapeutic relevance of T2M biology. Together, these findings highlight the ability of scMINER2 to overcome the curse of dimensionality, while preserving causal biological signals, reframing type-2 immunity as a hidden dimension of CAR-T persistence and durable cancer immunotherapy.
Other Authors: Jayadev Mavuluri, Song-Eun Lim, Darong Yang, Sheetal Bhatara, Xu Yang, Kevin Ye, Jai Mehta, Noemi Reyes, Mahesh Pujyan, Chun-Yang Lin, Qingfei Pan, Liang Ding, Koon-Kiu Yan, Terrence L. Geiger, Jiyang Yu
4) Synergy between CAR T and the endogenous immune system is key to a complete response in large B cell lymphoma
Chandler Gatenbee, Moffitt Cancer Center
Application of CD19 targeted Chimeric Antigen Receptor (CAR) T-cell therapy to large B cell lymphoma has yielded unprecedented clinical outcomes in a cancer largely resistant to conventional therapy. Under the assumption that CAR T cells and endogenous T cells interact, we hypothesize that the pre-treatment immune ecology, and the neoantigen burden it shapes, may be predictive of response to CAR T therapy. We take a multifaceted approach, combining mathematical modeling and whole exome sequencing to test our hypothesis. Our mathematical model, fit to clinical data, simulates tumor-immune coevolution before, during, and after CAR T therapy, allowing us to explore the relationship between pre-treatment neoantigen burden (NAB), the immune ecology, and outcome. Simulated NAB was then compared to patient pre-treatment NAB estimated from whole exome sequencing. Simulations suggest that complete response (CR) to CAR therapy is only expected when there is an appreciable pre-treatment endogenous antitumor immune response. This requires moderate to high expression of HLA and low immune evasion, which facilitates immune predation. Predation often increases diversity, here manifesting as high NAB. This model prediction was confirmed by our finding that observed HLA+ CR patients have significantly higher pre-treatment NAB than HLA+ relapsed/refractory (R/R) patients. Our findings suggest that effective CAR T therapy involves synergy with a pre-existing endogenous antitumor immune response, which leaves a signature of high NAB. While direct CAR T killing is HLA independent, our work suggests that HLA status, neoantigen burden, and the pre-treatment immune ecology, can be biomarkers for therapy decision making.
Other Authors: Mark Robertson-Tessi, Constanza Savid-Frontera, Bachisio Zicchedu, Venu V.G. Saralamma, Michael D. Jain, Jonathan H. Schatz, Francesco Maura, Frederick L. Locke, Alexander R.A. Anderson
5) Ex vivo models of tumor-immune-stromal interactions to study drug response in lung adenocarcinoma
Dina Hany, Stanford University
Lung adenocarcinoma (LUAD) is a leading cause of cancer-related deaths, with tumor heterogeneity and drug resistance as main therapeutic challenges. The tumor microenvironment (TME) primarily consists of malignant cells, stromal cells, and immune cells. These distinct cellular compartments respond differently to therapeutic agents based on their spatial organization and composition. In vitro or ex vivo models that preserve the native spatial organization of the TME would advance our understanding of the role of the TME in drug response. We are developing an in vitro multi-culture organoid system that includes patient-derived cancer cells, tumor-associated fibroblasts, and immune cells. Additionally, we are establishing an ex vivo culture system using precision-cut lung slices (PCLS) from freshly resected LUAD tissues. We employ advanced multimodal spatial-omic technologies (Xenium and Phenocycler Fusion) to characterize these model systems in terms of distinct spatial organization of these cellular compartments and their transcriptional states. Our recent computational framework enables us to integrate spatially resolved multi-omic data to identify functional gradients within the TME. This research elucidates the spatial dynamics of tumor-stroma-immune interactions and their impact on drug response within the TME for the overall goal to enhance personalized treatment strategies and improve patient outcomes.
Other Authors: Dina Hany, Anum Khan, Jacob Chang, Sylvia K. Plevritis
6) Phospho-proteomic Analysis of Chimeric Antigen Receptor Signaling in Natural Killer Cells
Yukiko Higa, University of Southern California
Chimeric antigen receptor (CAR)-engineered natural killer (NK) cells represent a promising “off-the-shelf” alternative to CAR-T immunotherapies for targeting hematologic malignancies. However, early signaling dynamics triggered by CAR engagement are not well understood, particularly across different cellular platforms. In this study, we employ Stable Isotope Labeling by Amino acids in Cell culture (SILAC) coupled with high-resolution mass spectrometry to comprehensively examine phosphorylation events in both primary donor-derived CAR-NK cells and an immortalized CAR-NK92 cell line, each engineered to target BCMA on multiple myeloma (MM) cells. We profiled phosphoproteomic changes at a 15 minute stimulation timepoint in co-culture, using a shared CAR construct and heavy labeled tumor cells to distinguish effector and target contributions. This design enables direct comparisons of signaling responses across cellular contexts—highlighting donor-to-donor variation in primary cells as well as intrinsic differences between immortalized and primary cell lines. Preliminary analyses demonstrate robust phosphorylation events enriched in known NK and CAR signaling pathways, and suggest both shared and divergent regulatory mechanisms between the two effector cell types. This work establishes a phosphoproteomics workflow for studying CAR-mediated signaling and offers insight into how cellular context influences the early activation landscape of CAR-NK therapies.
Other Authors: Lei Tian, Whitaker Cohn, Jianhua Yu, Stacey Finley, and Nicholas Graham
7) Homotypic fusion results in genome-doubled triple-negative breast cancer clones with altered phenotypes
Kennedy Howland, University of Texas at Austin
Intratumor heterogeneity drives tumor progression and therapeutic resistance, with mechanisms of whole-genome doubling (WGD) emerging as contributors. In triple-negative breast cancer (TNBC), spontaneous homotypic fusion events generate clones with stably doubled DNA content, yet the consequences of this state remain unclear. Here, we compared fusion-derived and control clones from the HCC1806 TNBC line to assess genomic stability and phenotypic divergence. Fusion clones displayed markedly higher ploidy, enlarged nuclear and cytoplasmic areas, and altered nuclear-to-cytoplasmic ratios relative to controls. Dose response curves generated for fusion and control clones in response to DNA damaging and microtubule-targeting chemotherapeutics do not show strong trends of innate resistance or susceptibility. Copy number variation (CNV) profiling revealed specific, distinct baseline CNV differences between fusion and control clones; and when repeated after 30 days no strong patterns of copy number reduction were observed, indicating relative genomic stability. Together, these findings suggest that homotypic fusion produces stable, genome-doubled cells with distinct morphological and proliferative phenotypes, contributing to tumor heterogeneity without promoting genomic instability or therapy resistance.
Other Authors: Andrea Gardner, Amy Brock
8) The Cancer Cell Map Initiative: 2025 Update
Trey Ideker, UC San Diego
In contrast to the substantial progress in decoding the genetic basis of cancer, systematic efforts to map cancer pathways are just beginning. The Cancer Cell Map Initiative (CCMI) has recently completed a multi-cancer protein–protein interaction map centered on the protein p53 measured in breast cancer, colon cancer, leukemia. We find that TP53 structural mutants interact robustly with chaperones and protein-folding proteins (e.g., DNAJA1), whereas TP53 DNA-binding mutants exhibit elevated metabolic signatures. These patterns illuminate the distinct interactions that differentiate different mutations of the same gene. To complement these efforts, we have completed a comprehensive map of synthetic lethal interactions between cancer genes and druggable targets across seven different tumor backgrounds, representing breast, lung and oropharyngeal cancers of heterogeneous oncogenic backgrounds. From combinatorial CRISPR knock outs, 1,805 synthetic lethal interactions were identified. Of these interactions, 226 translate to robust biomarkers of drug sensitivity, including frequent genetic alterations in the KDM5C/6A histone demethylases, which sensitize to inhibition of TIPARP (PARP7). Finally, we are using the above cancer network maps to guide interpretable precision oncology models called visible neural networks. Towards establishing these predictive models as physician-ready tools, we recently published a series of seven “Hallmarks of Predictive Oncology” (including Interpretability, Generalizability, and Fairness). This work includes a hallmarks-based scorecard and model checklist to guide both model creators and clinical evaluators, aiming to bridge the translational gap between machine learning innovations and clinical utility.
Other Authors: Trey Ideker, PhD, Professor, UC San Diego Marcus Kelley, PhD, Postdoc, UC San Diego Nevan Krogan, PhD, Professor, UC San Francisco Nadia Arang, PhD, Postdoc, UC San Francisco
9) Discovery of Novel Ecotypes for Immunotherapy Response Prediction in Non-small Cell Lung Cancer
Ilayda Ilerten, Stanford University
Non-small cell lung cancer (NSCLC) is the leading cause of cancer-related deaths globally. While immune checkpoint inhibitors (ICIs) have recently improved outcomes for patients with advanced stages of NSCLC, their effectiveness remains limited and existing FDA-approved biomarkers, have shown limited predictive accuracy for outcomes. In this study, we propose integrating the EcoTyper framework with bulk and single-cell sequencing data, alongside advanced single-cell imaging technologies, to identify novel biomarkers in NSCLC patients. EcoTyper is a deconvolution and machine learning-based framework that has been previously applied to thousands of tumor transcriptomes to uncover transcriptional state networks across various cell types within the tumor microenvironment. Preliminary findings indicate that lung-specific ecotype 6 (E6) and ecotype 2 (E2) surpass the progression free survival power when compared to the pan-cancer CE9 within an ICI-treated cohort of 110 LUAD patients. We further validated the ecotypes using single-cell RNA-seq data from nine public datasets from HTAN and TISCH2 covering over 250,000 cells, achieving up to 90% significant validation of cell states. We are currently in the process of spatially characterizing the ecotypes through spatial transcriptomics platforms. We anticipate that our approach will deepen our understanding of cell state diversity and biology in NSCLC, potentially leading to new biomarkers for predicting responses to ICI therapies.
Other Authors: Max Diehn, Andrew Gentles
10) Fast, flexible, learning-free organoid quantification and tracking with OrganoSeg2
Kevin Janes, University of Virginia
Organoids are routinely imaged by brightfield microscopy at low magnification, but these images are challenging to analyze quantitatively at scale. Given differences in organoid-culture format and image acquisition among research groups, there is a general need for versatile segmentation algorithms that refine for specific applications. Here, we introduce OrganoSeg2, an overhauled software that substantively advances the multi-window adaptive thresholding of its predecessor. OrganoSeg2 gives users access to additional segmentation parameters that were latent in OrganoSeg, and common operations are accelerated ~10-fold. Using data from six organoid types, we find that the generalized segmentation accuracy of OrganoSeg2 surpasses multiple alternatives, including segmenters based on deep learning. OrganoSeg2 adds longitudinal single-organoid tracking and multicolor fluorescence quantification, which we use to examine growth trajectories and radiotherapy responses in luminal breast cancer organoids. OrganoSeg2 is shared freely as installation packages for current users and source code for future developers (https://github.com/JanesLab/OrganoSeg2).
Other Authors: Cameron J. Wells, Najwa Labban, Shayna L. Showalter, Róża K. Przanowska
11) Clone-Specific Pathway and Program Inference in Low- to High-Grade IPMN Precursors of Pancreatic Cancer
Rachel Karchin, Johns Hopkins University
Intraductal papillary mucinous neoplasms (IPMNs) are common precursors to pancreatic cancer, yet the clonal and transcriptional transitions underlying progression from low- to high-grade lesions remain poorly understood. We collected 10 archival FFPE IPMN specimens containing both low- and high-grade regions and have begun in-depth analysis using Visium HD spatial transcriptomics and whole-exome sequencing. To interpret these data, we extended PictographPlus, originally developed to integrate bulk DNA and RNA sequencing for clonal evolution and clone-level expression inference. Because spatial transcriptomic data lack DNA information, we adapted PictographPlus to use copy-number clones inferred from spatial RNA profiles and identify altered pathways across clonal transitions using GSEA. CoGAPS identifies latent biological patterns from high-dimensional data, and ProjectR projects these patterns into independent datasets to enable robust cross-study and cross-platform comparisons. Here, we applied CoGAPS and ProjectR to transfer-learn patterns from a publicly-available bulk RNA-sequencing dataset containing IPMN, PDAC, and normal ductal epithelial samples. The patterns, representing cohort-level transcriptional programs, were projected onto our own IPMN Visium HD data to visualize their spatial localization. This dual framework of clone-aware pathway inference with PictographPlus and orthogonal projection of CoGAPS patterns with ProjectR will provide complementary insights into clonal architecture, transcriptional phenotypes, and microenvironmental interactions in early pancreatic tumorigenesis. Unlike existing approaches that rely solely on transcriptional profiling, our framework explicitly links spatial expression patterns to clonal evolution, establishing a novel paradigm for dissecting precancerous progression.
Other Authors: Jiaying Lai, Kathleen Noller, Prathima Nagendra, Adonis Dmello, Luciane Kagohara, Laura Wood, Elana Fertig
12) Dynamic modeling reveals that ErbB-targeted therapy with CDK4/6 inhibition prevents resistance and sustains response in ER+ breast cancer
Kimya Karimi, City of Hope
Cyclin-dependent kinase 4 and 6 inhibitors (CDK4/6i) have improved disease control in advanced estrogen receptor-positive (ER+) breast cancer, but acquired resistance remains a critical challenge limiting long-term efficacy. We previously found that patient cancer cells evade G1/S checkpoint control during treatment by boosting ErbB-driven growth signaling. Time-resolved transcriptomics of ER+ breast cancer cells exposed to ribociclib and afatinib were performed to reveal molecular dynamics driving resistance. This approach revealed that resistance arises through rapid rewiring of both cell cycle and cell death pathways. CDK4/6i monotherapy induced early activation of cyclin E/CDK2-E2F signaling, consistent with established resistance pathways. In contrast, the addition of afatinib suppressed both cyclin E/CDK2 at G1/S and cyclin B/CDK1 at G2/M, while markedly increasing caspase-dependent apoptosis. This combined blockade of cell cycle progression and enhancement of apoptosis enforced durable arrest and significantly prolonged response compared to monotherapy. These findings identify ErbB signaling as a central driver of CDK4/6i resistance and show that ErbB blockade complements CDK4/6 inhibition by simultaneously suppressing CDK2 and CDK1 pathways and driving apoptosis. More broadly, this study demonstrates that time-resolved modeling of dynamic signaling can uncover resistance mechanisms and guide treatment strategies, providing preclinical support for ErbB-targeted therapy as a strategy in patients progressing on CDK4/6i.
Other Authors: Jason I Griffiths, Eric F Medina, Elena Farmaki, Aritro Nath, Andrea H Bild
13) Massively parallel in situ generation and evaluation of cancer coding variants
Marcus Kelly, UCSD
The vast majority of cancer coding variants are of unknown function and lack direct experimental evaluation. While prime editing (PE) enables precise nucleotide edits, most amino-acid changes are difficult to induce with current PE guide RNA (pegRNA) designs. To expand the reach of PE technology to general amino-acid editing, we targeted diverse codons in 298 cancer genes with an expansive library of 21,437 pegRNAs, each coupled to an exogenous sensor of editing efficiency. ‘Flex’ pegRNAs, which induce multiple synonymous mutations within and outside the targeted codon, were often required to install the desired amino acid. Among the induced alterations were hundreds that affect cell fitness, recapitulating known oncogenic mutations and revealing previously unknown drivers. Using this PE dataset, we formulated a general-purpose transformer model for systematic design of amino-acid variant libraries. Together, these results enable general interrogation of the coding mutations found in cancer genomes.
Other Authors: -
14) Ex vivo models of tumor-immune-stromal interactions to study drug response in lung adenocarcinoma
Anum Khan, Stanford University
Lung adenocarcinoma (LUAD) is a leading cause of cancer-related deaths, with tumor heterogeneity and drug resistance as main therapeutic challenges. The tumor microenvironment (TME) primarily consists of malignant cells, stromal cells, and immune cells. These distinct cellular compartments respond differently to therapeutic agents based on their spatial organization and composition. In vitro or ex vivo models that preserve the native spatial organization of the TME would advance our understanding of the role of the TME in drug response. We are developing an in vitro multi-culture organoid system that includes patient-derived cancer cells, tumor-associated fibroblasts, and immune cells. Additionally, we are establishing an ex vivo culture system using precision-cut lung slices (PCLS) from freshly resected LUAD tissues. We employ advanced multimodal spatial-omic technologies (Xenium and Phenocycler Fusion) to characterize these model systems in terms of distinct spatial organization of these cellular compartments and their transcriptional states. Our recent computational framework enables us to integrate spatially resolved multi-omic data to identify functional gradients within the TME. This research elucidates the spatial dynamics of tumor-stroma-immune interactions and their impact on drug response within the TME for the overall goal to enhance personalized treatment strategies and improve patient outcomes.
Other Authors: Dina Hany, Anum Khan, Jacob Chang, Sylvia K. Plevritis
15) Systems Genetics and Integrative Genomics Reveal Insights into Breast Cancer from Genetically Engineered Mouse Models (GEMMs)
Minjeong Kim, University of Tennessee Health Science Center
Triple-negative breast cancer (TNBC) lacks effective targeted therapies due to its molecular and phenotypic heterogeneity. We developed a novel preclinical model, the BXD-Breast Cancer (BXD-BC), derived from C57BL/6, DBA/2J, and FVB/N [SV-40-C3(1)-Tag (“C3Tag”)] backgrounds. This model provides a genetically controlled platform to investigate TNBC biology, displaying diverse tumor phenotypes. In this study, we aim to define underlying genetics and biomarkers of tumor onset and progression and prioritize conserved candidates. Previously, we identified quantitative trait loci (QTL) and cis-expression QTL (cis-eQTL) associated with tumor phenotypes in 28 BXD-BC strains. We now investigate genotyped single nucleotide variants and genome-wide alterations. RNA-seq analyses revealed BXD-BC molecular subtypes i that mirror the heterogeneity of human TNBC, including prognostic signatures. Whole-genome analysis uncovered copy number alterations on several chromosomes, notably chromosome 6 near Kras, the integration site of the C3(1)-Tag transgene, and deletions on chromosomes 2, 7, and 10. Integrative analyses demonstrated that T-antigen copy number and Kras amplification are associated with distinct gene expression clusters. Systems genetics analysis further identified strain-specific single nucleotide variants (SNVs) organized into haplotype blocks, potentially highlighting inherited CNAs across BXD strains that influence gene expression and tumor-associated phenotypes. Further analyses will identify selectivity on genes that promote tumor onset or progression, encompassing systems genetics and genomics. BXD-BC is a valuable platform to uncover TNBC drivers and guide biomarker development and targeted therapeutic strategies.
Other Authors: Jeremiah R. Holt, Xiaobei Zhao, Laura M. Sipe, David G. Ashbrook, Robert W. Williams, Liza Makowski, D. Neil Hayes.
16) Clone-Specific Pathway and Program Inference in Low- to High-Grade IPMN Precursors of Pancreatic Cancer
Jiaying Lai, Johns Hopkins University
Intraductal papillary mucinous neoplasms (IPMNs) are common precursors to pancreatic cancer, yet the clonal and transcriptional transitions underlying progression from low- to high-grade lesions remain poorly understood. We collected 10 archival FFPE IPMN specimens containing both low- and high-grade regions and have begun in-depth analysis using Visium HD spatial transcriptomics and whole-exome sequencing. To interpret these data, we extended PictographPlus, originally developed to integrate bulk DNA and RNA sequencing for clonal evolution and clone-level expression inference. Because spatial transcriptomic data lack DNA information, we adapted PictographPlus to use copy-number clones inferred from spatial RNA profiles and identify altered pathways across clonal transitions using GSEA. CoGAPS identifies latent biological patterns from high-dimensional data, and ProjectR projects these patterns into independent datasets to enable robust cross-study and cross-platform comparisons. Here, we applied CoGAPS and ProjectR to transfer-learn patterns from a publicly-available bulk RNA-sequencing dataset containing IPMN, PDAC, and normal ductal epithelial samples. The patterns, representing cohort-level transcriptional programs, were projected onto our own IPMN Visium HD data to visualize their spatial localization. This dual framework of clone-aware pathway inference with PictographPlus and orthogonal projection of CoGAPS patterns with ProjectR will provide complementary insights into clonal architecture, transcriptional phenotypes, and microenvironmental interactions in early pancreatic tumorigenesis. Unlike existing approaches that rely solely on transcriptional profiling, our framework explicitly links spatial expression patterns to clonal evolution, establishing a novel paradigm for dissecting precancerous progression.
Other Authors: Kathleen Noller, Prathima Nagendra, Adonis Dmello, Luciane Kagohara, Laura Wood, Elana Fertig, Rachel Karchin
17) Functional and spatial mapping of proteins to decipher cellular heterogeneity
William Leineweber, Stanford University
Most proteins localize to multiple cellular locations, yet their canonical function is typically associated with only one location. This lack of understanding is compounded in cancer, where many disease-associated mutations cause proteins to mislocalize. Improved mapping of subcellular protein localizations is needed to understand resulting morphological and functional impacts. Towards this end, we have used image-based spatial proteomic approaches to develop analytic and functional technologies to address these challenges. 1) We have created a suite of AI models, collectively dubbed “SubCell”, that capture key cellular features that are useful for downstream biological analyses. We demonstrate that SubCell can reliably predict cell types and protein localizations from single-cell images, identify morphological shifts in triple negative breast cancer resulting from vorinostat and paclitaxel drug treatments, and reconstruct structural gene ontologies. 2) We also performed analysis of cell shapes and associated relationships with organelles, pathways, and individual proteins. We found shared modes of variation across diverse cell types, conserved organelle spatial relationships within, but not between, cell lines, and cell-shape dependent differences in metabolic proteins in the G2 phase of the cell cycle. 3) Lastly, we have developed a novel imaging platform to link live-cell functional behaviors to the proteomic states of cancer cells. Cell migration and proliferation are measured in single-cells, followed by an end-point readout using highly-multiplexed immunofluorescence imaging. From these measurements we can identify cancer cell subpopulations that differ proteomically and functionally.
Other Authors: Trang Le, Emma Lundberg, Ankit Gupta, Jan Hansen
18) Nutrient Availability Shapes Cooperative Interactions Between Triple-Negative Breast Cancer Clones
Sarah Meng, UT Austin
Triple-negative breast cancer (TNBC) comprises 15–20% of cases and remains the most clinically challenging subtype due to the lack of targeted therapies and high heterogeneity. Previous studies have focused on defining molecular subtypes of TNBC to better understand the molecular heterogeneity and guide treatment. However, the metabolic heterogeneity and its impact on clonal interactions and tumor progression is less understood. To investigate this, we generated single-cell-derived clonal populations from the MDA-MB-231 TNBC cell line and cultured these clones under high (4.5 g/L) and low (1 g/L) glucose conditions. Transcriptomic profiling revealed that clones separated into two clusters distinguished by ESAM expression. Each clone exhibits distinct metabolic specialization based on the ESAM expression, showing complementary metabolic profiles between clones. We then assessed clonal interactions by growing populations in monoculture and coculture. Cocultures pairing ESAM+ and ESAM- clones exhibited cooperative interactions that significantly enhanced growth compared to monocultures. These findings indicate that metabolic specialization drives nutrient-dependent cooperation between clonal populations. Our study highlights the importance of metabolic heterogeneity and cooperation in TNBC progression and suggests that disrupting nutrient-dependent metabolic interactions may be a novel therapeutic approach to target tumor heterogeneity and inhibit tumor growth.
Other Authors: Amy Brock
19) Retatrutide alleviates metabolic dysfunction and delays tumor latency in obesity-associated breast cancer
Naveed Pervaiz, The University of Tennessee Health Science Center, Memphis, TN
Breast cancer (BC) is second leading cause of cancer death among women in the US. Obesity increases risk of multiple cancers, including BC. Previous studies demonstrated that GLP-1 receptor agonists promote weight loss, improve metabolic dysfunction, reduce risk and progression in pancreatic and lung cancer, but effects on spontaneous BC were unknown. This study aims to assess the impact of a novel triple incretin receptor agonist, retatrutide (RETA), on BC risk and tumor progression in mice with obesity-associated metabolic dysfunction. Female FVB C3(1)-Tantigen mice, a genetically engineered mouse model (GEMM) of BC, were fed a high-fat diet from 8 weeks of age. Overweight mice were subjected to RETA as weight loss intervention or weight matched (WM) controls. At 11 weeks, mice received subcutaneous RETA (15nmol/kg) or vehicle every other day. One group was sacrificed at 15 weeks to investigate preneoplastic lesions; the second group was allowed to proceed to endpoint to examine tumor onset, progression, and survival. RETA and WM induced significant weight loss (~15–20%) within 2–3 weeks, followed by weight stabilization, with reduced circulating leptin concentration. RETA and WM reduced fasting blood glucose, decreased gonadal white adipose tissue, and mammary fat. RETA and WM moderately reduced muscle mass in preneoplasia group. RETA delayed gastric emptying revealed by higher cecal contents compared to vehicle control and WM. In IACUC endpoint group, RETA significantly delayed tumor latency. Altogether, results demonstrated that pharmacological weight loss suppressed early tumor progression-despite being in a transgenically-driven GEMM and improved outcomes in obesity-associated BC.
Other Authors: Sandesh J Marathe, Zereque Powell, Logan G. McGrath, Zeid T. Mustafa, Liza Makowski
20) Mechanisms linking Drp1 and mitochondrial fission to tumor growth in KRas-driven colon cancer.
Daniel Phipps, University of Virginia
Colorectal cancer (CRC) develops in a complex metabolic environment. Approximately 40% of CRC patients carry a deleterious mutation in KRAS, a potent driver of tumor cell metabolism. Oncogenic KRas signaling promotes dynamic structural changes in mitochondrial networks via ERK-mediated activation of the mitochondrial fission GTPase Drp1. CRC cell lines exhibit a highly fragmented mitochondrial phenotype, but the mechanisms that link aberrant mitochondrial fission to tumor cell growth and metabolism are poorly understood. We hypothesize that KRas-driven Drp1 activation initiates specific metabolic changes to promote tumorigenesis within the complex energetic environment of the colon. These specific changes in metabolism, however, could create vulnerabilities that could potentially be exploited therapeutically. To address this hypothesis, we are utilizing an integrated systems biology approach combining inducible Drp1 knockouts in vitro, computational metabolic models derived from transcriptomic data, and a genetically engineered mouse model of Drp1 loss in spontaneously developing CRC tumors. Our initial analyses reveal that Drp1 is required for low density colony formation of human CRC cells. Furthermore, acute Drp1 deletion leads to changes in expression of genes involved in innate immune signaling and differentiation status, providing potential insights into the mechanisms through which Ras-induced mitochondrial fission promotes cellular proliferation. Ultimately, we seek to exploit these mitochondrial fission-dependent phenotypes to identify novel therapeutic targets that will limit the growth and metastatic potential of tumors that arise in the colon.
Other Authors: William Shao, Jennifer A. Kashatus, Tiffany A. Melhuish, David Wotton, Jason A. Papin, David F. Kashatus
21) Evolutionary immunotherapy in NSCLC: identifying optimal dosing strategies in adoptive cell therapies using agent-based modeling
Sandhya Prabhakaran, Moffitt Cancer Center
BACKGROUND: TIL therapy is an emerging immunotherapy where activated T cells are injected into the patient. This therapy can fail due to tumor-induced immunosuppression, for example via the PDL1/PD1 axis. PDL1 expression has been studied and can increase under IFNg, released by activated T cells, but PD-L1 relaxation is not understood. We hypothesize that there may be better strategies of therapy delivery that maximize tumor kill while minimizing immune suppression. We investigate these complex PD-L1 driven dynamics through a unique combination of in vitro studies and mathematical modeling. METHODS: We developed a hybrid agent-based model (ABM). The agents in the ABM are tumor cells having variable PD-L1 expression, and immune cells that secrete IFNg and can kill tumor cells. Tumor cells can be randomly or cluster seeded. IFNg is modeled as either a binary well-mixed pulse (to simulate the in vitro experimental setup), or as a diffusible via a reaction-diffusion process (to simulate the tumor-immune cell interactions in an in vivo spatial manner). We model TIL therapy as an immune cell pulse, given at regular or irregular intervals and with different quantities of cells. RESULTS: Our ABM is calibrated to capture in vitro data dynamics. Under certain combinations of spatial configurations of the tumor and intermittent immune dosing schedules, we observe the presence of ‘sweet spots’ where tumor extinction is possible and immune exhaustion is avoided. This indicates that an appropriate pulsing of TIL therapy may lead to better overall immune efficacy than a bolus injection or continuous immunotherapy, by preventing sustained suppression that increases immune cell exhaustion, despite the potential for tumor regrowth between pulses. Our novel findings can potentially benefit clinical cancer research by giving multiple insights related to the tumor extinction, equilibrium and escape phenomena.
Other Authors: Mark Robertson-Tessi, Kimberly Luddy, Rafael Bravo, Taylor M. Bursell, Julian Pineiro, Jeffrey West1, Megan Johnson, Jhanelle E. Gray, Amer A. Beg, Scott Antonia, Robert A. Gatenby and Alexander R. A. Anderson
22) Proteome-wide prediction of interactions between structured domains and peptide motifs reveals functionally coherent sub-networks
Aakash Saha, Columbia University
Protein–protein interactions mediated by short linear motifs (SLiMs) are essential for cellular signaling and regulation yet remain underrepresented in current interactome maps due to their transient and low-affinity nature. Here, we present PrePPI-SLiM, a proteome-scale computational pipeline that integrates domain–motif interaction data from the ELM database with structural domain annotations and sequence-derived features such as disorder and conservation. Utilizing a naïve Bayes classifier and a likelihood ratio-based scoring framework, PrePPI-SLiM systematically evaluates all possible protein pairs within a proteome to identify interactions plausibly mediated by domain–motif recognition. Applied to the human and yeast proteomes, PrePPI-SLiM predicted over 21 million PPepIs in humans and 1.7 million in yeast, including nearly 4 million and 430,000 interactions, respectively, not found in the current PrePPI database containing domain-domain interactions. At a stringent false positive rate (FPR ≤ 0.005), the pipeline identified 114,935 high-confidence intreractions in humans spanning 59 Pfam domains, and 8,580 PPepIs in yeast spanning 32 Pfam domains, demonstrating broad coverage of peptide recognition domains across species. To assess structural plausibility, we modeled high-confidence interactions using AF3Complex and validated them against experimentally resolved receptor–peptide complexes from the Propedia database. The strong agreement between predicted and known structures highlights the biological relevance of these interactions. Furthermore, clustering of the high-confidence PrePPI-SLiM interactome yields functionally coherent networks that reveal mechanistic insights into cellular processes. Altogether, PrePPI-SLiM provides a scalable and statistically robust framework for uncovering peptide-mediated interactions that are often missed by existing experimental and computational methods.
Other Authors: Caroline Valez, Diana Murray, Barry Honig
23) Metformin reduces T2D-related tumor aggressiveness by modifying plasma exosome miRNA profile
Michael Seen, Boston University
Uncontrolled diabetes is associated with poorer outcomes among prostate cancer patients. In previous studies, it has been found that miRNAs carried in plasma exosomes play a role in driving more aggressive prostate tumor behavior. We hypothesize that metformin may modify the miRNA payload of plasma exosomes, thereby partially mitigating these tumor enhancing effects. Platelet-free plasma was obtained from insulin resistant mice who were either treated with metformin or untreated. Exosomes were isolated using size-exclusion chromatography. miRNAs were isolated from exosomes, then transfected into DU145 cells using lipofectamine RNAiMax. Migratory ability was measured by means of transwell migration assay. Cells transfected with miRNAs obtained from the plasma exosomes of mice who were on metformin demonstrated significantly reduced migratory ability compared with cells transfected with miRNAs from the plasma exosomes of untreated insulin resistant mice. The miRNA payload of plasma exosomes appears to be altered by treatment with metformin in a way that partially mitigates the pro-migratory effects of plasma exosomal miRNA on prostate tumor cells.
Other Authors: Naser Jafari, Andrew Chen, Manohar Kolla, Isabella R. Pompa, Yuhan Qiu, Rebecca Yu, Pablo Llevenes, Christina S. Ennis, Joakin Mori, Kiana Mahdaviani, Meredith Halpin, Gretchen A. Gignac, Christopher M. Heaphy, Stefano Monti, Gerald V. Denis
24) Context-specific transcriptomic response to acute loss of Drp1 in colorectal cancer
William Shao, University of Virginia
Colorectal cancer (CRC) is a global health burden, ranking as the 2nd leading cause of cancer-related deaths worldwide. A key driver of CRC progression is the dysregulation of mitochondrial fission, leading to a fragmented mitochondrial network which promotes tumor growth. However, the precise mechanisms linking mitochondrial fragmentation to tumor growth remain unclear. To investigate this mechanism, we created an inducible knockout (KO) system for Drp1, the primary mediator of mitochondrial fission. Using this system, we compared the transcriptomic response to Drp1 KO in two distinct CRC cell lines: HCT116, which exhibits a fragmented mitochondrial network, and DLD-1, which exhibits a more fused network. RNA-sequencing was performed pre- and post-Drp1-KO to identify effects on downstream pathways. Pathway enrichment analysis and genome-scale metabolic modeling revealed divergent responses to loss of mitochondrial fission. HCT116 cells exhibited an innate immune response and shift towards a more epithelial state, as well as upregulation of autophagy and peroxisomal fatty acid oxidation. In contrast, DLD-1 cells activated an ER stress response and upregulated fatty acid synthesis. These results suggest that the role of mitochondrial fission may be highly dependent on the cell’s metabolic and mitochondrial state. We are currently working to experimentally validate these hypotheses, with the goal of identifying context-specific vulnerabilities to exploit as novel treatments for CRC.
Other Authors: Daniel N. Phipps, Jennifer A. Kashatus, Tiffany A. Melhuish, David Wotton, David F. Kashatus, Jason A. Papin
25) Conception-calibrated pediatric tumor mitotic clocks
Darryl Shibata, USC School of Medicine
Molecular clocks can reconstruct tumorigenesis, but their calibration is limited by uncertainty in cancer ages. Pediatric cancers simplify this problem because age ranges are narrowly bounded by conception and a minimum of ~30 divisions. We developed mitotic clocks from rapidly fluctuating CpG (fCpG) DNA methylation on the X chromosome in male cancers, applying a binary Markov model to estimate mitotic age and epimutation rates. Across acute lymphoblastic leukemia, acute myeloid leukemia, neuroblastoma, and embryonal brain tumors, modeled and observed methylation data were highly concordant, yielding epimutation rates of ~10-3 per division. The clocks also resolved remission dynamics, inferring relapses seeded by small numbers of variably dormant residual cancer cells. By exploiting the unique age constraints of pediatric tumors, calibrated fCpG clocks provide a quantitative framework to reveal otherwise hidden features of human tumor evolution.
Other Authors: Jeremiah John
26) Plasma-Derived Exosomes from Type 2 Diabetic Patients increase Tumor Aggressiveness and Chemoresistance in ER+ Breast Cancer
Austyn Smithback, Boston University School of Medicine
Type 2 diabetes (T2D) is a rapidly growing global health crisis, affecting over 500 million people worldwide. Women with T2D have a higher risk of developing breast cancer of all subtypes and tend to experience poorer cancer outcomes, yet metabolic status remains underrepresented in treatment guidelines. Exosomes, small extracellular vesicles carrying bioactive cargo such as microRNAs, are emerging as key mediators of communication between metabolic disease and cancer. However, the role of T2D plasma-derived exosomes in breast cancer progression remains poorly understood. Using MCF-7 cells, a well-established ER+ breast cancer model, we examined the impact of exosomes from T2D patient plasma on chemoresistance and tumor aggressiveness. Compared to untreated controls, T2D exosomal miRNA-treated cells displayed significantly greater survival after vincristine exposure and reduced apoptotic response, consistent with enhanced chemoresistance. Furthermore, T2D exosome-treated cells exhibited a 2.7-fold increase in EMT marker expression and a 1.9-fold increase in migratory capacity, as evidenced by accelerated wound closure, compared to non-diabetic controls, indicating a phenotypic shift toward enhanced invasiveness. Overall, our results support T2D-derived exosomes as key drivers of chemoresistance and tumor aggressiveness in ER+ breast cancer, an exciting finding that advances our understanding of chemoresistance in patients with metabolic disease. These insights may inform the development of novel therapeutic strategies to overcome drug resistance and improve treatment efficacy in this high-risk patient population. This work was supported by U01 CA243004 from the Cancer Systems Biology Consortium.
Other Authors: Michael Seen, Dr. Gerald Denis; Boston University School of Medicine
27) A spatial reaction-diffusion model of mitotic signaling suggests how breast tumors generate Chromosomal Instability.
Todd Stukenberg, University of Virginia, School of Medicine
Cancer cells often lower the fidelity of mitosis to drive tumor progression which is known as Chromosomal Instability (CIN), but the cellular changes that underlie CIN have remained mysterious. We previously published that breast tumors generate CIN by combining loss of P53 function with the dysregulation of the mitotic transcription factors FoxM1 and MybL2. These transcription factors coregulate most of the mitotic transcriptional program and many of their targets are also overexpressed in highly aneuploid tumors. However, it is unclear which of these proteins contributes to CIN. Since dysregulation of Tyrosine kinase signaling drives tumor progression, we hypothesized that the overexpression of FoxM1 and MybL2 dysregulates mitotic signaling to cause CIN. We provide evidence that cells in breast tumors dysregulate the transcription of the mitotic signaling kinases including Aurora B, Plk1, Bub1 and Mps1 as well as their regulators to generate CIN. We built a highly constrained spatial reaction diffusion model of mitotic signaling networks (combining spindle checkpoint, CPC localization and kinetochore-microtubule resolution pathways) to test this hypothesis. Emergent properties of the network include the concentration of the Chromosomal Passenger Complex (CPC) to the inner centromere, as well as kinetochore-microtubule dependent changes to CPC localization and spindle checkpoint signaling. Our future goal will be to use patient RNA levels of genes in the network as inputs to the model and predict the amount of aneuploidy in each patient’s tumor, which would providing evidence that tumors affect mitotic signaling to generate CIN. Overall, our model provides strong evidence for a network architecture that controls the events of the mitotic spindle. There are a number of surprises including the finding that a single signaling network coordinates the regulation of of multiple mitotic events.
Other Authors: Catalina Alvorez Yela, Sarah M. Groves, Monserrat Gerardo, Jasraj Raguwanshi, Kevin Janes
28) Deep-learning–driven tumor microenvironment profiling improves immunotherapy response prediction in ccRCC
Ruohan Wang, Stanford University
Clear cell renal cell carcinoma (ccRCC) exhibits extensive cellular heterogeneity within the tumor microenvironment (TME), which shapes immune infiltration patterns and affects patient response to immunotherapy. However, current biomarkers remain insufficient to capture this complexity and predict treatment response. We developed a ccRCC-specific Ecotyper model using 614 bulk RNA-seq samples from TCGA to define 56 distinct cell states across immune, stromal, and malignant populations, grouped into 11 ecotypes. Applying this model to 98 treatment-naïve samples from a clinical CRC immunotherapy cohort, we found that ecotype E2 was significantly associated with favorable response to immune checkpoint inhibitors (ICIs). Spatial transcriptomics further confirmed the coordinated localization of E2 cell states within tumor tissue. To elucidate the regulatory mechanisms underlying ecotype behavior, we integrated single-cell and spatial transcriptomics data to construct ecotype-specific gene regulatory networks, capturing intra- and inter-cellular communication. We then trained a graph attention network (GAT) using a pan-cancer immunotherapy dataset (n=1,934) and fine-tuned it on the ccRCC cohort. The GAT model achieved the highest predictive accuracy with AUC scores of 0.786-1.0, compared to conventional methods and biomarkers. Attention weights from the GAT model identified key immune-modulatory genes, including GZMA, NKG7, and CLEC9A, as well as novel candidates, such as GAS7 and LRRC25, offering mechanistic insights into treatment responsiveness. Together, our findings provide a systems-level framework to dissect TME architecture and enhance personalized immunotherapy strategies in ccRCC.
Other Authors: Andrew Gentles
29) Using spatial multi-omic assays to explore sequential therapy approaches for melanoma
Chong Xia, Institute for Systems Biology
The CSBC Research Center – Spatiotemporal Tumor Analytics (ST-Analytics) – investigates the mechanisms behind the observed durability of a combination of immune checkpoint blockade (ICB) followed by a MAPK inhibitor (MAPKi) for treating melanoma. Currently we are building models for phenotyping H&E regions through integration with spatial chromatin accessibility assessed via DBiT ATAC-seq (deterministic barcode in tissue sequencing for transposase-accessible chromatin) and working to integrate H&E, CODEX, and mIHC images from adjacent sections of murine tumors through whole slide image registration. After creating tiled image data, mIHC and CODEX tiles are processed to provide machine learning targets; i.e. whether a particular cell type is observed. Machine-vision deep learning models are being trained to detect the presence of immune cells, starting with CD8 T cells. These predictions will be correlated to previously predicted areas of lymphocyte infiltration producing a well calibrated model that can be applied in H&Es integrated with DBiT data. Future directions include the generation of relevant spatiotemporal features that allow the design of agent-based models (ABMs) to represent the dynamics of tumor immune microenvironment which can inform experimental combination treatments of targeted inhibitors and immunotherapy in cancer models.
Other Authors: Juho Kim, Shuo Wang, Jianjun Jiang, Yin Tang, Sarah Li, David L Gibbs, Boris Aguilar, Heber L Rocha, Paul Macklin, Claudia M Ludwig, Ilya Shmulevich, Wei Wei, Vésteinn Thorsson, James R Heath
30) Systems biology analysis identifies the synthetic lethality between NEK5 and VHL in clear cell renal cell carcinoma
Zhen Xie, St Jude Children's Research Hospital
Inactivation of the von Hippel–Lindau (VHL) tumor suppressor gene appears in over 80% of clear cell renal cell carcinomas (ccRCC), presenting a major therapeutic vulnerability. We aimed to exploit VHL synthetic lethality (SL), where simultaneous perturbation of VHL and its SL partner induces cancer cell death, to develop strategies for targeting VHL-null diseases. To reconcile the discrepancy between RNA expression and functional activity, we used SJARACNe (Algorithm for the Reconstruction of Accurate Cellular Network), to reconstruct a ccRCC-specific gene-gene interactome from TCGA-KIRC and obtained high-fidelity regulons for 5,724 transcription factors and 20,764 signaling proteins. This approach enabled the accurate inference of hidden driver activities using NetBID2 (data-driven network-based Bayesian inference of drivers). We then applied this framework to 274 RNA-seq profiles from 84 ccRCC patients (containing 57 VHL null and 27 VHL WT) enrolled in the UTSW SPORE cohort to systematically search for hidden VHL-SL drivers. We identified NEK5, an understudied kinase as a SLG in VHL-null ccRCC. Functional validation in cell lines and mouse models confirmed that NEK5 loss suppressed VHL-deficient tumor growth, an effect reversed by restoring VHL expression. Mechanistically, NEK5 knockdown perturbed mitochondrial function, increased ROS production that inhibits ccRCC tumor growth, which aligned to the AP-MS–derived NEK5 interactome enriched for proteins involved in energy metabolism and stress response. Furthermore, NEK5-targeting PROTAC prototypes demonstrated potent, selective inhibition of VHL-null tumor growth in vitro. In summary, we report NEK5 as a mitochondrial regulator that is synthetic lethal with VHL, highlighting a novel therapeutic avenue for ccRCC.
Other Authors: Cheng Zhang, Shanshan Bradford, Jamie Jarusiewicz, Gisele Nishiguchi, Kaiwen Yu, Qingfei Pan, Junmin Peng, Qing Zhang, Jiyang Yu
31) Spotiphy enables single-cell spatial whole transcriptomics across an entire section
Jiyuan Yang, St. Jude Children's Research Hospital
Spatial transcriptomics (ST) has advanced our understanding of tissue regionalization by enabling the visualization of gene expression across an intact tissue section, but current platforms face the dilemma between genome-wide coverage and single-cell resolution. Here we present Spotiphy, a novel algorithm that generates inferred single-cell RNA profiles (iscRNA data) of all cells to achieve spatially resolved whole-slide transcriptomic profiling. Extensive benchmarking demonstrated Spotiphy's superiority against thirteen other methods. In evaluations using matched scRNA-seq, Visium, Xenium, CosMx, and immunohistochemistry (IHC) datasets for Alzheimer’s Disease (AD) mouse brains, Spotiphy delivers the most precise cell-type proportions including rare cell populations. Analysis using the iscRNA data reveals novel regional specifications of astrocytes. It distinguishes unique sub-populations of disease-associated microglia (DAM) located in different brain regions. Spotiphy identifies alterations in tumor-tumor microenvironment (TME) interactions across distinct spatial domains of human breast cancer samples. For the first time, Spotiphy tackles information loss in the non-capture area. Spotiphy recovers the expression profiles of non-tagged nuclei in Slide-tags. Spotiphy enables single-cell spatial whole transcriptomics across the entire section, significantly increasing the information density of ST data. It offers innovative spatial systems biology analysis pipelines for exploring cellular organization, tissue heterogeneity, and the function of complex biological systems.
Other Authors: Chenghuan Liu, Masayuki Umeda, Shanshan Bradford, Qiqi Jin, SongEun Lim, Jing Ma, Sheetal Bhatara, Tamara Westover, Jeffery Klco, Jiyang Yu
32) A hidden-driver approach identifies synergistic brain-penetrant drug combinations for high-risk medulloblastoma and beyond
Jiyang Yu, St. Jude Children's Research Hospital
Targeted strategies are urgently needed but remain elusive for high-risk Group 3 medulloblastoma, a metastatic, recurrence-prone subtype that responds poorly to current non-specific therapies. To accelerate discovery of targeted therapies for this malignancy, we developed SINBA (Synergy Inference by Data-driven Network-Based Bayesian Analysis), a systems biology framework that identifies covert disease drivers and then nominates synergistic drug pairs that target them. The inclusion of a medulloblastoma-specific gene network and a curated drug-gene interaction database allows SINBA to prioritize drug combinations that are both synergistic and blood-brain-barrier-permeable. Screening of 320 drug combinations derived from the in silico analysis of over 10,000 driver pairs produced 19 synergistic pairs; of these, a combination of MEK inhibitors and the p38 inhibitor regorafenib was particularly efficacious. This combination suppressed G3 tumor progression in mice, with therapeutic benefits further enhanced by low-dose radiation. Single-cell RNA sequencing found early UBC progenitor-like tumor cells to be especially sensitive to the combination treatment, which remodels the tumor microenvironment to promote antitumor immune responses. These results demonstrate SINBA’s great potential as a drug discovery platform, clinically actionable G3 MB strategy, and precision oncology framework.
Other Authors: Jingjing Liu, Xu Yang, Mingrui Zhu, Xinran Dong, Honglei Zhou, Brandon Bianski, Barbara M. Jonchere, Wenwei Lin, Xiang Fu, Abigail Wang,Ruilin Jiang, Lei Yang, Burgess B. Freeman III, Taosheng Chen, Giles W. Robinson, Martine F. Roussel, Thomas E. Merchant, Amar Gajjar