Hi @AWGall
Today’s Horizons in Biosciences and Informatics Seminar series (HBISS) featured Dr. Andreas Keller (Saarland University and the Helmholtz Institute for Pharmaceutical Research Saarland) with his co-author Friederike Grandke, on what longer spaceflight does to the whole mammal at the molecular level, one small RNA at a time, across 13 organs. 
Here is the link to view the recording if you were unable to join. [RECORDING LINK]
This was the Rodent Research-8 / RRRM-1 talk many of you have been waiting for. The 31 RR-8 omics datasets on OSDR are open and ready for analysis.
What Andreas covered
Andreas’s group studies aging as a systemic process, starting with non-coding RNAs (miRNAs in particular) and tracing how those small regulators reshape genes, cells, and whole organs. The longer goal is comprehensive AI models that map this behavior across organs and organisms, since evolutionary conservation is where the transferable signal lives. This study extends an organism-wide framework from his Stanford work with the Wyss-Coray lab, which produced body-wide ncRNA atlases of aging and parabiosis (miR-29c-3p was the strongest miRNA restored by rejuvenation) and a 15-region brain miRNA atlas pinning microglial miR-155-5p as a brain-aging driver.
The RR-8 design analyzed 686 small RNA samples from female mice across 13 solid organs at 3 and 8 months, after at least 3 weeks on the ISS, against Earth controls. Three groups separated the variables: Flight (FL), Habitat Ground Control (HGC, ISS-matched housing including CO2, temperature, humidity), and Vivarium Ground Control (VGC, standard housing). Two protocols ran in parallel, Live Animal Return (LAR, the focus) and Immediate Termination on the ISS.
Disentangling spaceflight from housing. This was the part Andreas flagged as most surprising. Comparing FL vs HGC, FL vs VGC, and HGC vs VGC separated true spaceflight effects from the effect of being housed differently on Earth. The split is strongly organ-specific. Brain, kidney, and mesenteric adipose were shaped mostly by housing. Gut/gonadal adipose (GAT), spleen, subcutaneous adipose (SCAT), thymus, liver, and pancreas were driven by spaceflight, strongest in GAT, spleen, and SCAT. The brain is the cautionary case: much of its miRNA change tracks with Earth housing differences, not microgravity, which is exactly why control design matters in these missions.
Tissue-specific, orchestrated, and linked to genes. Most spaceflight-deregulated miRNAs are tissue-specific, with only a small systemic set. The MIR-17/92 and MIR-1/133 families drive distinct changes through specific gene targeting, and a hypothesis-free GSEA showed coordinated action concentrated in SCAT, spleen, and pancreas, clustering around tissue remodeling and ECM, developmental change, and DNA damage repair, along a Fat-Liver-Pancreas axis. Because the same mice also have single-cell gene expression data, the team matched deregulated miRNAs to the genes they target, answering why genes change, not just which. Brown adipose showed the most miRNA-targeted gene deregulation, and many top targets are transcription factors that amplify a small signal broadly.
Aging versus spaceflight. Age-dependent changes are real but smaller in magnitude than the age-independent spaceflight changes, with the strongest age effect in the thymus and age-dependent effects channeling through MIR-8, MIR-154, and MIR-15 in diaphragm, pancreas, and MAT. Compared against the Tabula Muris Senis aging cohort, the data do not show a strong acceleration of aging. Andreas was careful here: organs age differently, not faster. If this holds in humans, even older adults could remain viable candidates for long-duration missions such as a trip to Mars.
A note before you analyze the scRNA-seq. After the miRNA paper published, a colleague in Japan flagged a sample switch affecting a subset of thymus and spleen samples. This affects the single-cell gene expression data, not the miRNA data. The datasets were reannotated. If you are working with the RR-8 scRNA-seq data, be aware of the this issue explained in this post.
Key discussion highlights
A genuinely robust Q&A, forty minutes of questions after a twenty-minute talk.
- Lynn Harrison asked whether deregulated miRNAs tie to CO2 and metabolism. Andreas: the control cohorts differ on too many factors to deconvolve CO2 from these mice alone.
- Grant Goodman @gwgoodman asked about muscle-specific miRNA dysregulation with aging. Andreas: yes, organ-specific signals in heart and skeletal muscle, with strong follow-up interest in muscle and fat.
- Jian Gong @jgong proposed a fluid-shift and pressure hypothesis tying brain and retina to astronaut vision changes. Andreas called it fascinating, but the metadata to test it was not delivered with the dataset.
- Ryan Scott @rtscott2001 asked about environmental telemetry access. The team did not receive it, and all flight mice shared a cage. I flagged that the Environmental Data App currently has only radiation data for RR-8 and will follow up with the data team.
- Lane Christensen @lchristenson (RR-20 PI) asked about mitochondrial-function miRNAs. Andreas: yes, even more so in the single-cell data, with energy metabolism in the fat tissues leading.
- Marshleen Yadav @Marshleen asked about exosomal miRNAs. Andreas: no plasma material was available, and the classical exosomal signatures did not appear, fitting how tissue-specific these miRNAs are.
- Nilufar Ali @nilufarali asked, for the Brain AWG aging subgroup, whether the brain ages faster and whether the team can share aging miRNA gene sets for GSEA. Andreas was enthusiastic on both. The brain is effectively 15 organs with very different regional aging patterns, and he wants organ experts in the community to take his per-organ miRNA and gene lists and dig in. He called it a community project.
- Nicholas Brereton @nicholas.brereton noted this is a natural fit for sub-AWG community projects, with env data plus RadLab plus RNA-seq as the integration to aim for.
- On foundation models, @rtscott2001 asked about fine-tuning scGPT for spaceflight. Andreas: enough data exists, but spaceflight confounders (the RR-8 food bar issue is a good example) make him prefer results-level integration here, and classical ML like XGBoost still wins many molecular benchmarks.
- Friederike Grandke (co-author on the RR-8 Nature Comms miRNA paper) addressed the food bar question: no molecular outlier effects, but it shortened the terminal-group flight, making that group less comparable to the live-animal-return group.
The repeated theme from both speakers: the per-organ miRNA and gene lists are open, the team wants collaborators, and the right next step for the AWG is to pick organs and projects and reach out. The Brain AWG aging subgroup is an obvious start. See the master table of open sub-group projects.
Next HBISS
June 25, 2026: Kexin Huang on AI agents building AI biology models (Biomni, Phylo.bio). 
Watch Forum-Space for time and details.
All links from the chat and presentation
- RR-8 study, Grandke et al. 2026, Nature Communications (open access)
- Spatio-temporal brain miRNA atlas, Engel et al. 2025, Nature Communications
- Aging and heterochronic parabiosis ncRNA atlas, Wagner et al. 2024, Nature Biotechnology
- scGPT single-cell foundation model, Cui et al. 2024, Nature Methods
- 31 RR-8 / RRRM-1 datasets on OSDR (Forum-Space)
- RR-8 scRNA-seq annotation update (read before analyzing)
- Chan Zuckerberg CELLxGENE Discover
- OSDR Data Visualization Portal
- Environmental Data App (EDA)
- Andreas Keller Google Scholar
- Master table of open AWG sub-group projects
- HBISS series
- AWG Forum-Space About