2025 Hyper Recent •CC0 1.0 Universal

This work is dedicated to the public domain. No rights reserved.

Access Preprint From Server
May 9th, 2025
Version: 1
National Cancer Institute
cancer biology
biorxiv

Deep learning inference of cell type-specific gene expression from breast tumor histopathology

Wang, A. T.Open in Google Scholar•Dhruba, S. R.Open in Google Scholar•Wang, K.Open in Google Scholar•Campagnolo, E. M.Open in Google Scholar•Shulman, E. D.Open in Google Scholar•Ruppin, E.Open in Google Scholar

Cell type-specific gene expression from single-cell RNA sequencing (RNA-seq) is valuable for breast cancer precision oncology but available cohorts are still limited due to its high cost. Deconvolution methods infer cell type-specific expression from bulk RNA-seq at a lower cost, yet expenses and processing time of bulk RNA-seq are also non-negligible and limit their application too. To address these limitations, we developed SLIDE-EX (SLide-based Inference of DEconvolved gene EXpression), a deep-learning tool that predicts cell type-specific gene expression and abundances directly from routine breast cancer histopathology whole slide images (WSIs), using deconvolved bulk RNA-seq data as training labels. Trained on the TCGA-breast cohort and tested in cross validation and on an independent cohort of 160 cases, SLIDE-EX robustly infers the expression of thousands of genes across 9 distinct cell types, performing best for cancer associated fibroblasts and cancer cells. The abundance of these two cell types could also be robustly predicted, together with that of myeloid cells. The robustly predicted genes reflect key biological functions of their respective cell types. From a translational perspective, the inferred cell type specific expression profiles predict chemotherapy response more accurately than models based on direct prediction from the slides or from the inferred bulk expression in two independent cohorts. Going forward, SLIDE-EX is a generic approach that opens up possibilities for rapid, cost-effective cell type-specific gene expression inference in potentially any cancer type, further democratizing the characterization of the tumor microenvironment.

Similar Papers

biorxiv
Fri May 09 2025
BRAFV600E-Driven Lung Tumorigenesis Requires Ligand-Mediated Activation of ERBB Receptor Signaling
Secretion of ligands of the human epidermal growth factor (EGFR) family of receptors or erythroblastic leukemia viral oncogene family (ERBB1-4) is a feature common to many cancer cells. However, our understanding of the role of autocrine ligands in the aberrant behavior of cancer remains incomplete. Here we demonstrate that, in numerous preclinical models of lung tumorigenesis, BRAFV600E signaling...
Dacheux, M.
•
Wu, M.-J.
•
Scherzer, M. T.
•
Nillson, M.
...•
Vaishnavi, A.
biorxiv
Fri May 09 2025
Improved Tumor Blood Flow Enhances the Abscopal Effect: Preclinical Assessment in Mice Treated with Combined Radiation and PD-1 Blockade Therapy
The abscopal effect, where localized radiation therapy induces regression of distant metastatic lesions through immune activation, shows promise for treating metastatic cancer but occurs inconsistently. Here we demonstrate that tumor perfusion critically influences systemic immune responses to combination therapy with radiation and PD-1 blockade. Using multimodal imaging including DCE-MRI, EPR oxi...
Yamashita, K.
•
Brender, J. R.
•
Saida, Y.
•
Otowa, Y.
...•
Kishimoto, S.
biorxiv
Fri May 09 2025
A polyomavirus-positive Merkel cell carcinoma mouse model supports a unified cancer origin
The Germ Cell Theory, grounded in developmental biology, has enabled significant advances in understanding and treating germ cell tumors (GCTs), whereas somatic cancer research, dominated by the Somatic Mutation Theory (SMT), has stagnated and demands a paradigm shift. Accumulating evidence challenges the traditional separation between GCTs and somatic cancers. Polyomavirus-positive Merkel cell ca...
yang, w.
•
contente, s.
•
rahman, s.
biorxiv
Fri May 09 2025
HIF1α and HIF2α mRNA Expression in Oral Squamous Cell Carcinoma does not correlate with protein expression
Background: Oral cancer is one of the common cancers among men, its presentation is usually late. Tumor growth leads to hypoxia, and HIF1 and HIF2 are important regulators of hypoxia induced pathways, inducing angiogenesis and tumor growth. This study compares the expression levels of HIF1 and HIF2 mRNA in OSCC tissues to adjacent normal mucosa. Methods: RNA was isolated from 85 OSCC tissues and a...
Pandey, M.
•
Singh, P.
•
Nath, G.
•
Kumar, D.
•
Kar, A. G.
biorxiv
Fri May 09 2025
Differential DNA damage response to WRN inhibition identifies a targetable vulnerability in ARID1A-mutated cancers
ARID1A, a key subunit of the SWI/SNF chromatin remodeling complex, is frequently mutated in cancers. However, effective clinical treatments for patients with this mutation are limited, highlighting a demand for new therapeutic strategies. Here, we establish Werner syndrome ATP-dependent helicase (WRN) as a critical vulnerability target in ARID1A-mutated cancers. Upon genetic and pharmacological in...
Kim, J.
•
Oh, J.
•
Jang, D.
•
Shin, S.
...•
Cho, S.-Y.
biorxiv
Fri May 09 2025
Bone marrow lymphocyte dynamics during chemotherapy in pediatric acute myeloid leukemia
A better understanding of lymphocyte dynamics during current treatment regimens in pediatric AML is urgently needed to understand whether the application of bispecific T-cell-engagers (BiTEs) during periods of low tumor burden could be a viable treatment strategy. In this study, we found that induction 1, comprising mitoxantrone-etoposide-cytarabine in nearly all patients (as part of the NOPHO-DBH...
Koedijk, J. B.
•
Barneh, F.
•
Meesters-Ensing, J. I.
•
van Tuil, M.
...•
Heidenreich, O.
biorxiv
Fri May 09 2025
Multiomic profiling of hypoxic glioblastoma stem cells reveals expansion of subpopulations with distinct epigenetic and CNV profiles
Glioblastoma is characterized by extensive intratumoral heterogeneity driven by dynamic microenvironmental cues such as hypoxia. While transcriptional and epigenetic variability have been separately linked to hypoxia responses, the integrated impact of hypoxia on gene regulation and clonal architecture in glioblastoma stem cells (GSCs) remains poorly defined. We applied single-nucleus multiomics n...
Corseri, A.
•
Moore, T. W.
•
Szczepanski, N.
•
Hwang, H.
...•
Tapinos, N.
biorxiv
Fri May 09 2025
Single cell spatial transcriptomics identifies coordinated cellular programs associated with good prognosis in microsatellite stable colorectal cancer
It has been reported that microsatellite instable (MSI) tumors in colorectal cancer (CRC) exhibit stronger anti-tumor responses than microsatellite stable (MSS) tumors. Analyzing scRNA-seq data from 185 CRC patients we found that immune, structural, and cancer cells in MSS tumors with high numbers of tumor-infiltrating CD8 T cells and macrophages (TAMs; CD8hiTAMhi) and CD8lowTAMhi tumors were enri...
Karlsen, V. T.
•
Majid, U.
•
Lindholm, H. T.
•
Umu, S. U.
...•
Jahnsen, F. L.
biorxiv
Fri May 09 2025
SRSF3 knockdown-induced cellular senescence as a possible therapeutic strategy for non-small cell lung cancer
Tyrosine kinase (TK) inhibitors improve clinical outcomes in non-small cell lung cancer (NSCLC) with targetable mutations. However, such NSCLC cases only consist of about 50% in the western populations. This study, for the first time in NSCLC cells including those without a targetable TK mutation, explores a tumor-suppressive activity of siRNA knockdown of a splicing factor SRSF3, which was report...
Nakamichi, S.
•
von Muhlinen, N.
•
Yamada, L.
•
Melamed, J. R.
...•
Harris, C. C.