Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/191064

Curated compendium of human transcriptional biomarker data

Golightly, N. P.·Bischoff, A. I.·Bell, A.·Hollingsworth, P. D.·Piccolo, S. R.
View original source ↗Export metadata ↓

Abstract

Genome-wide transcriptional profiles provide broad insights into cellular activity. One important use of such data isto identify relationships between transcription levels and patient outcomes. These translational insights can guide the development of biomarkers for predicting outcomes in clinical settings. Over the past decades, data from many translational-biomarker studies have been deposited in public repositories, enabling other scientists to reuse the data in follow-up studies. However, data-reuse efforts require considerable time and expertise because transcriptional data are generated using heterogeneous profiling technologies, preprocessed using diverse normalization procedures, and annotated in non-standard ways. To address this problem, we curated a compendium of 45 translational-biomarker datasets from the public domain. To increase the datas utility, we reprocessed the raw expression data using a standard computational pipeline and standardized the clinical annotations in a fully reproducible manner (see osf.io/ssk3t). We believe these data will be particularly useful to researchers seeking to validate gene-level findings or to perform benchmarking studies--for example, to compare and optimize machine-learning algorithms ability to predict biomedical outcomes.

Source connections

Explore related subjects

cancer biologyPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Golightly, N. P., Bischoff, A. I., Bell, A., Hollingsworth, P. D., Piccolo, S. R.. 2017-09-19. Curated compendium of human transcriptional biomarker data. https://doi.org/10.1101/191064

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

TROP2 and immune gene expression by African ancestry in triple-negative breast cancer

Purpose: TROP2 is a treatment target in triple-negative breast cancer (TNBC). TACSTD2 encodes TROP2, while PDCD1 encodes the immune protein PD-1. We investigated whether these genes differ by race or African genetic ancestry and whether the reported link between higher TACSTD2 and lower immune-gene expression partly reflects the mixture of cells in tumor samples. Methods: We analyzed TNBC tumors from TCGA (n = 107) and GSE268851 (n = 253; ancestry data for 229), together with spatial, normal-breast, and protein datasets. Regression models adjusted for age, tumor stage, and a score based on epithelial gene expression. We compared TACSTD2 with other epithelial markers using paired bootstrap analysis. Results: TACSTD2 was detected in nearly all tumors, with no significant differences by race or ancestry. Higher African ancestry was linked to higher PDCD1 expression in TCGA (P = 0.00039), but not in GSE268851. In GSE268851, higher TACSTD2 remained associated with lower CD8A, GZMB, and cytotoxic score after epithelial adjustment (all q < 0.003). These associations were stronger than those for EPCAM; the CD8A association was also stronger than that for KRT8. TACSTD2 was 1.55-fold higher in TNBC than in normal breast and correlated with TROP2 protein (Spearman rho = 0.74). Conclusion: TACSTD2 was widely expressed in the TNBC groups studied, while the ancestry-PDCD1 association differed between cohorts. Epithelial gene expression did not fully explain some negative TACSTD2-immune associations, and some were stronger than those for other epithelial markers such as EPCAM.

cancer biology↗
bioRxiv2026

Quantifying subcellular organization from localization to co-occurrence with InSpace

Cell identity and function depend not only on molecular abundance, but also on how molecules are organized within cells. Although spatial transcriptomics now resolves individual RNA molecules within cells, routine computational workflows reduce these measurements to cell-by-gene count matrices, retaining abundance while discarding information about intracellular localization and molecular co-occurrence. This loss of spatial information creates a need for scalable methods that preserve subcellular detail and enable comparisons across diverse cellular morphologies. Therefore, we present InSpace, a scalable framework for quantifying these complementary dimensions of subcellular organization. InSpace-R maps transcripts onto a continuous radial coordinate anchored to nuclear and cell boundaries, while InSpace-C measures distance-dependent co-occurrence or depletion of molecular pairs. Simulations and orthogonal biochemical references support accurate measurements, with scalability demonstrated on datasets containing 275 million transcripts. RNA-RNA and RNA-DNA analyses recover organelle-linked neighbourhoods and associations between nascent transcripts and their genomic origins. Analysis of B and plasma cells across 58 tumour datasets spanning 12 cancer types reveals that reduced MHC-II transcript abundance accompanies more peripheral localization and stronger mutual co-occurrence, separating these transcripts from a program enriched for ER-targeted proteins. InSpace makes intracellular organization accessible alongside abundance, enabling discovery of spatial programs that distinguish cellular states and generate hypotheses about their functional specialization.

cancer biology↗
bioRxiv2026

Age-based shifts in multicellular community organization and abundance in HPV-Negative Head and Neck Cancer

Aging represents a fundamental driver of cancer risk and progression, yet the mechanisms by which aging reshapes the tumor microenvironment remain incompletely understood. To interrogate how aging alters cellular and microenvironmental organization in cancer, we developed an integrated single-cell atlas of HPV-negative head and neck squamous cell carcinoma (HNSCC) encompassing 73 patients aged 18-91 years. Using a multi-layered computational framework combining NMF-based module discovery and partial correlation network analysis, we reveal that age-associated differences in tumor composition are driven primarily by selective shifts in cell type abundance and organization rather than widespread age-dependent transcriptional rewiring. We identify an age-enriched cellular community centered on basal-like squamous epithelial states linked to epithelial-mesenchymal transition, matrix-associated myofibroblast-like CAFs, and FOLR2+ macrophages, which secrete immunosuppressive and ECM-remodeling factors. Cross-dataset in-silico validation via TCGA transcriptomes, spatial transcriptomics, and mouse tumor models confirms the correlation of these cell types across patients, specific ligand-receptor usage, and increased proportions of epithelial and stromal cells in aged mouse models. Among epithelial cell-intrinsic age-associated transcriptional changes, midkine (MDK) emerges as a rare but robust tumor-derived signal linked to stromal remodeling, immunosuppression, and invasive epithelial programs. These findings reveal conserved mechanisms of age-dependent cellular reorganization in the tumor microenvironment and identify MDK and associated stromal-remodeling pathways as potential intervention points for age-stratified cancer therapy.

cancer biology↗