Single-Cell Sequencing Extends Tumor Analysis to Fixed Clinical Samples

Single-cell RNA sequencing has become a powerful tool for examining the cellular diversity of tumors, but its reach inside clinical trials is often limited by the way patient samples are collected, stored and processed. Fresh tissue is difficult to preserve outside specialized laboratories, while formalin-fixed and paraffin-embedded material—the standard format for many clinical studies—can compromise conventional cell-profiling workflows. Researchers at Roche have now benchmarked a newer fixed-tissue approach and found that the 10X Genomics GEM-X Flex Gene Expression assay, combined with a “chop-fix” preparation protocol, can generate comprehensive transcriptomic data from frozen tumors and archived biopsy blocks collected under routine clinical-trial conditions.

The work, published in BMC Genomics, addresses a practical bottleneck rather than a single biological question. Single-cell sequencing can distinguish rare cell populations, characterize states that blur together in bulk tissue and reveal how tumor cells interact with immune and stromal components. Yet those benefits depend on obtaining viable or otherwise processable cells with minimal degradation. Clinical trial specimens often travel between hospitals, central laboratories and analytical sites, and delays or differences in preservation can impose major logistical constraints. A workflow compatible with fixed tissue could therefore make longitudinal molecular analysis more feasible and preserve samples that might otherwise be unsuitable for single-cell studies.

To evaluate the approach systematically, the team compared fixed-tissue and nuclei-based RNA-sequencing workflows in solid-tumor samples collected and preserved using procedures commonly encountered in clinical trials. The comparison included both frozen tumor tissue and formalin-fixed and paraffin-embedded, or FFPE, blocks. Rather than testing an isolated laboratory protocol, the study placed the GEM-X Flex assay against the established GEM-X Universal Gene Expression solution under conditions representative of real-world trial specimens. This design allowed the investigators to assess both technical data quality and the extent to which biologically relevant signals could be recovered.

Technical performance was evaluated through several related measures, including the number of genes and unique molecular identifiers, or UMIs, detected in each cell. UMIs are molecular tags that help quantify captured RNA, making their recovery an important indicator of transcriptomic sensitivity. The GEM-X Flex workflow combined with chop-fix preprocessing produced higher gene and UMI detection than the standard GEM-X Universal Gene Expression solution in both frozen and FFPE samples. It also yielded a slight reduction in ambient RNA background, the unwanted signal from RNA released outside cells that can dilute or distort cell-specific profiles. Together, these results indicate that the newer workflow did not merely recover a different number of cells; it preserved more molecular information and modestly improved the cleanliness of the resulting data.

The preparation stage is central to that improvement. Chop-fix mechanically processes tissue into small, manageable fragments before fixation, helping researchers create a suspension from which individual nuclei can be isolated and profiled. Nuclear RNA sequencing avoids the need to keep whole cells intact, an advantage for solid tumors and clinical specimens that may be fixed, stored or shipped under conditions incompatible with fresh-tissue workflows. The study’s results suggest that carefully combining tissue homogenization with the Flex assay can recover a richer transcriptomic profile than the established fixed-sample solution, while still retaining the practical advantages of nuclei-based analysis.

Beyond aggregate sequencing metrics, the investigators examined whether the workflows recovered specific biological signals and distinguished cell states that conventional approaches left less clear. The chop-fix and GEM-X Flex combination improved the capture of relevant biological signals and resolved cellular states that were less distinct with the comparison method. This matters because a higher number of detected genes does not by itself guarantee a more informative cell atlas. Useful single-cell data must preserve differences among tumor, immune and stromal populations, including subtle states that may be linked to treatment response, resistance or disease progression. The benchmark therefore suggests that the protocol improved both the amount and the biological interpretability of the data.

The samples were part of the infrastructure surrounding clinical trial NCT04338685, registered on ClinicalTrials.gov. The study was approved by the applicable independent ethics committees and regulatory authorities and conducted in accordance with the Declaration of Helsinki, ICH Good Clinical Practice E6(R1), the European Union Clinical Trials Directive and local requirements. Written informed consent was obtained from participants before eligibility assessment. The research was funded by F. Hoffmann-La Roche, and the authors disclosed affiliations with Roche entities, including employment or company shareholdings for several contributors. The use of trial-linked material illustrates the practical setting the workflow was designed to support: molecular analysis of specimens collected for therapeutic research rather than specimens prepared specifically for a sequencing laboratory.

Fixed-tissue single-cell sequencing could widen the questions that clinical investigators can ask, but the study does not make archival samples interchangeable with newly collected tissue. Formalin fixation, paraffin embedding, storage history and mechanical processing can all influence molecular recovery. The benchmark’s importance lies in demonstrating that an established single-cell platform can perform well with an adapted preparation method, not that all fixed specimens will yield equivalent results. Researchers would still need to validate sample quality, examine platform-specific quality-control metrics and interpret differences in sensitivity when comparing fixed-tissue datasets with fresh or frozen alternatives. Such controls are especially important when subtle cell states are central to a trial’s biological hypothesis.

Overall, the results position the 10X Genomics GEM-X Flex Gene Expression assay with chop-fix preprocessing as a robust option for transcriptomic analysis of both frozen tissue and FFPE clinical biopsies. Its main advantage is not a claim of viral discovery or a dramatic single benchmark result, but the conversion of a persistent workflow problem into a testable technical solution. By increasing gene and UMI detection, modestly lowering ambient RNA background and improving the representation of distinct cellular states, the approach could make single-cell analysis more compatible with the way clinical-trial specimens are actually collected and preserved. If adopted carefully, fixed-tissue profiling may give researchers a more practical route from patient biopsy to a high-resolution view of tumor biology, while keeping molecular analysis closer to the clinical evidence gathered during treatment.

Subject of Research: Performance of fixed-tissue single-cell RNA sequencing for clinical-trial tumor samples

Article Title: Performance of the 10X Genomics flex single-cell sequencing assay and its application to overcome challenges in clinical trial samples

Article References: Antoniolli, M., Servera, L. A., Paetzold, K., Schmeing, S., Yong, C., Nassiri, S., Hüsser, T., Cannarile, M. A., Bacac, M., Yángüez, E., & Dettling, S. (2026). Performance of the 10X Genomics flex single-cell sequencing assay and its application to overcome challenges in clinical trial samples. BMC Genomics. https://doi.org/10.1186/s12864-026-13476-2

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13476-2

Keywords: single-cell RNA sequencing, 10X Genomics, GEM-X Flex, fixed tissue, FFPE, clinical trials, tumor microenvironment, RNA sequencing, chop-fix protocol, solid tumors, Performance, Genomics