As global investment in brain atlas projects exceeds $200 million, a new editorial in Medicine Discovery argues that atlas scale alone cannot deliver clinical benefit — and calls for functional testing, harmonized metadata, and transparent reporting
Zhujiang Hospital of Southern Medical University
Brain science is entering an era in which single-cell, single-nucleus, spatial, and multimodal omics can be interpreted within a shared biological framework. Yet a new editorial published in Medicine Discovery argues that increasing atlas scale or resolution does not by itself establish biological meaning — a gap with direct clinical consequences, as more than 92% of drugs that show promise in animal models never reach human patients.
The editorial, Frontiers in Brain Science and Multi-Omics: From Atlas Construction to Functional Testing and Translational Research , is written by Peng Luo of the Department of Oncology at Zhujiang Hospital of Southern Medical University and Hongbo Guo of the hospital’s Neurosurgery Center. It synthesizes recent multi-omics studies of the brain and sets out three complementary strategies for testing candidate regulatory relationships: dynamic comparisons, joint measurement of multiple molecular layers, and targeted perturbation experiments.
“Progress toward clinically useful discovery will require complete metadata, transparent analytical procedures, independent corroboration, and clear distinction between association, prioritization, and functional testing,” the authors write.
From association to mechanism
The editorial illustrates the gap between prioritization and mechanism with a concrete example. A postmortem study of Alzheimer’s disease found that reduced epigenomic stability was associated with pathological progression, whereas its relative preservation was associated with cognitive resilience. This finding — drawn from postmortem brain tissue — shows how multi-omics can connect molecular states with disease progression. But the authors caution that such associations require functional testing before they can be interpreted as causal mechanisms.
The atlas landscape itself is already broad. Population-scale resources such as brainSCOPE integrate gene expression, chromatin accessibility, genetic variation, and cell–cell communication in the adult prefrontal cortex, covering more than 2.8 million nuclei from 388 individuals. Cross-modal and whole-brain resources extend coverage further: an adult mouse primary motor cortex atlas resolves more than 56 neuronal cell types and supports cross-platform cell-type alignment, while an adult whole-brain three-dimensional atlas profiles 117 dissected regions. MAPbrain compiles data across multiple brain regions and developmental stages in humans and nonhuman primates for cross-species comparison.
According to the editorial, these resources extend atlas coverage across stages, individuals, regions, and species — but scale or resolution alone cannot determine the functional relevance of a candidate regulatory relationship. Prioritizing a candidate gene is not the same as establishing a disease mechanism. For example, MultiVINE-seq can link genome-wide association study risk variants to human brain vascular, perivascular, and immune cell types and prioritize candidate genes, but mechanistic interpretations still require in vivo or perturbation-based testing.
Three complementary strategies
To move beyond prioritization, the authors highlight three complementary strategies. Dynamic comparisons capture how responses vary over time and across brain regions and cell types. Joint measurements read several molecular layers at once: 3DRAM-seq jointly profiles the three-dimensional genome, chromatin accessibility, DNA methylation, and RNA expression in human cortical organoids, while the Spatial Augmented Multiomics Interface (Sami) integrates the metabolome, lipidome, and glycome within a single mouse brain section. Functional testing then closes the loop: in human brain organoids, multiplexed CRISPR perturbations tested how GLI3 and HES4/5 regulate telencephalic fate, and a study of malformations of cortical development combined single-nucleus RNA sequencing of resected patient brain tissue with experiments in the embryonic mouse brain to link candidate somatic mutations to specific developmental cell populations.
The editorial then traces how these approaches are being applied to disease and to neuroengineering — and why the biological context of each study matters. In donors aged 0–3 years with trisomy 21, the dorsolateral prefrontal cortex showed transcriptional and chromatin alterations across cell types, together with disruption of programs related to synaptic function, myelination, and inflammation. In a cohort of 16 patients with pediatric high-grade glioma, comparison of diagnostic and post-treatment samples showed spatial co-occurrence of proneural tumor cells and non-neoplastic oligodendrocytes, together with altered myeloid cell states. And in a mouse cortical implantation model using non-functional silicon probes, molecular changes related to immune responses, neuronal health, and autophagy varied with time and distance from the implant, while mRNA and protein changes were discordant.
A central message of the editorial is that these findings come from very different biological materials — postnatal human brain tissue, postmortem brain tissue, diagnostic and post-treatment tumor samples, and an animal implantation model — and should not be interpreted as arising from a common mechanism. The same caution applies more broadly: findings from postmortem tissue, patient samples, organoids, and animal models should not be treated as interchangeable.
What progress requires
The authors also set out what progress will require in practice. Consistent sample-processing procedures, complete metadata, and clear reporting standards are needed so that cross-cohort comparisons can specify sample sources, data-processing procedures, and analytical conditions, and so that main conclusions can be independently verified. Key findings should be corroborated in independent cohorts, clinical samples, or functional models, with data, code, and analytical parameters fully documented.
The editorial stresses that functional models and clinical samples serve different purposes: functional models can test candidate relationships, whereas clinical samples can show whether the corresponding phenomena occur in humans.
The editorial concludes that teams spanning basic science, clinical medicine, engineering, and data science should focus on clearly defined, testable medical questions with practical relevance, integrating multiscale measurement, functional testing, and assessment of potential applications from the outset. The aim, the authors suggest, is not to accumulate an ever larger volume of brain measurements, but to transform increasingly complex data into reproducible biological insight and experimentally testable opportunities for diagnosis, therapeutic development, and neuroengineering.
Article information
Frontiers in Brain Science and Multi-Omics: From Atlas Construction to Functional Testing and Translational Research .
Medicine Discovery , 2026.
Original article link: https://www.sciencedirect.com/science/article/pii/S3117928226000019
DOI: 10.1016/j.medd.2026.100001
About the journal
Medicine Discovery (ISSN: 3117-9282) is a peer-reviewed, international, open-access journal published by Zhujiang Hospital of Southern Medical University in collaboration with KeAi Publishing and distributed on the ScienceDirect platform. The journal is dedicated to publishing high-quality translational medical research that bridges mechanistic discovery and clinical implementation, and covers five sections spanning mechanisms and systems medicine, translational and clinical research, digital medicine and data-driven healthcare, therapeutic innovation and biomedical engineering, and methodology, standards, and implementation science.
Website: https://www.sciencedirect.com/journal/medicine-discovery
Submit: https://www.editorialmanager.com/medd/default2.aspx
Commentary/editorial
Not applicable
Frontiers in brain science and multi-omics: From atlas construction to functional testing and translational research
Hongbo Guo is Editor-in-Chief of Medicine Discovery and was excluded from editorial decision-making on this article.