Daily Summary for 2026-10-20

Introduction to the show: ident: Genomics Radio. Generated commentary on the latest computational biology and genomics papers.

Ines: It's the twentieth of October, twenty twenty-six, and this is the day's research.

Marcus: 6 new papers came out today.

Ines: I'm Ines.

Marcus: And I'm Marcus, here with Yuki.

Ines: We'll take the day in one pass, then pull out the papers we're staying with.

Marcus: Start us off with Direct RNA sequence design under codon constraints using expressive tensor-based secondary structure models.

The summary: Ines: Welcome to the twentieth of October, twenty twenty six.

Marcus: This research explores building more trustworthy regulatory candidate discovery.

Yuki: Requiring different computational systems to agree on a target gene before calling it a driver is key.

Ines: Orchestra: Corroboration-Based Regulatory Candidate Discovery via Composed Bioinformatics MCP Agents is detailed in that paper.

Marcus: Relying on just one piece of evidence can lead to false positives when identifying key disease regulators.

Yuki: The main idea behind Orchestra is combining two separate bioinformatics servers into a single workflow.

Ines: RegNetAgents infers regulatory network topology within that system.

Marcus: CASCADE gathers independent functional evidence like LINCS knockdown or super-enhancer status.

Yuki: So they must agree on the candidate regulator before it's called a driver?

Ines: That ensures corroboration across different computational systems.

Ines: Testing with two out of four pieces of evidence improves prediction for known driver status in BRCA and COAD tumors.

Marcus: That shows an odds ratio of 3.15 from papers 2609.05496.

Yuki: Multiple independent checks confirm a gene is important, making it more likely to be a true driver than one source alone.

Ines: This corroboration effect holds for other cancer types like STAD, where the odds ratio was 6.71 in papers 2609.05496.

Marcus: The benefit is more modest in BRCA/COAD compared to STAD results from papers 2609.05496.

Yuki: Researchers removed duplicate data and the permutation test confirmed that removing them reproduced original findings from RegNetAgents papers 2609.05496.

Ines: So, removing duplicates did not change the results for RegNetAgents papers 2609.05496.

Marcus: Agreed, and the permutation test confirmed that removal reproduced original findings from RegNetAgents papers 2609.05496.

Lucky paper: 2604.19718: Ines: We're looking at Direct RNA sequence design under codon constraints using expressive tensor-based secondary structure models.

Marcus: This paper tackles the huge problem of designing RNA sequences for proteins when you have to stick to specific codons.

Yuki: It develops a direct and efficient algorithm that samples sequences from a Boltzmann distribution defined by both the codon sequence and a detailed secondary structure free energy model.

Ines: So it moves beyond just optimizing stability; it integrates the codon constraint directly into the sampling process.

Marcus: The key contribution is using dynamic programming with tensor-based models of secondary structure to do this exactly.

Yuki: They use concepts like tensor trains and traces of matrix products to formulate hard and soft constraints.

Ines: And they show how they combine these two approaches by replacing the standard folding function, chi, with a generalized version,.

Marcus: This allows them to compute the free energy q(phi) exactly through the expression TrS

Q zero k T k,n(Q k+1,n-one): .

Yuki: They can then sample sequences phi directly from D(psi) or D lambda(psi) using Algorithm S4 point 1.

Ines: The complexity is stated as O(n three), but they say the empirical complexity is often lower than that for many problem sizes.

Marcus: They also give you exact computation costs for the marginal probabilities, which are O(n three) for the base pair probability matrix and O(n) for codon frequencies.

Yuki: The paper computes the equilibrium base pair probability matrix, P psi, with an O(n three) cost.

Ines: And they show that computing the marginal codon probability, p psi i(c), can be done in just O(n) complexity for any given index i.

Marcus: The paper proves that these algorithms incur no approximations and only a roughly four times cost prefactor compared to analyzing a fixed RNA sequence.

Yuki: This framework offers a principled way to do codon optimization based on dynamic programming and tensor modeling, which is expected to be a new paradigm for computational sequence design.

Lucky paper: 2511.10807: Ines: So we're moving on to a paper that looks at how activity fronts can stabilize vulnerable systems, titled Invading activity fronts or near-critical clusters may stabilize vulnerable excitable systems against stochastic trajectories terminating in an absorbing state.

Marcus: This is about two finite regions, one robust and one vulnerable to extinction, but they get stabilized by coupling them diffusively.

Yuki: The core mechanism involves inducing activity influx across the interfaces from the stable patch into the vulnerable region.

Ines: It sounds like this paper is looking at how physical structure can prevent a system from going extinct when it's small.

Marcus: Yeah, but it distinguishes between two types of extinction transitions in nonlinear stochastic systems, which is important context for that.

Yuki: They separate genuine active-to-absorbing state phase transitions, which require taking the thermodynamic limit N to infinity first.

Ines: But then there are stochastic extinction events in finite small systems, triggered by temporal trajectories reaching an absorbing state.

Marcus: And those stochastic extinctions are highly sensitive to the system size N, and often to the control parameters you set.

Yuki: The main finding is that diffusively coupling these two regions induces activity influx across the interfaces from the stable patch.

Ines: If those invasion fronts enter the inactive, excitable subsystem, they reinvigorate it and keep it immune to reaching the absorbing state.

Marcus: They used a simple model with two competing species, predator and prey, where one patch had carrying capacity K = one and the other K = three.

Yuki: So the smaller left patch in a simulation with two coupled systems was in a stable active predator-prey coexistence state with persistent radially propagating activity waves.

Ines: And that larger right subsystem, which was prone to fluctuation-driven total population extinction, got stabilized by those invasion fronts colliding.

Marcus: They found the basic requirements are just that extinction times depend sensitively on system size, and the source region stays stable during coupling.

Yuki: And crucially, the absorbing region needs to be excitable, meaning it can sustain traveling invasion fronts.

Lucky paper: 2503.17418: Ines: We're looking at the Application of Single-cell Deep Learning in Elucidating the Mapping Relationship Between Visceral and Body Surface Inflammatory Patterns, which is paper number three of three this time.

Marcus: This study uses single-cell deep learning to map how inflammation inside organs relates to inflammation on the surface.

Yuki: The core finding here is identifying a fascia-visceral inflammatory crosstalk pattern or FVICP in response to Salmonella dysentery.

Ines: That systemic impact of visceral inflammation is what really stood out in this work.

Marcus: They looked at the transition of expression profile in subcutaneous fascia cells after they were induced with Salmonella dysentery.

Yuki: The deep learning framework was able to show that this visceral inflammation influences cell states located distally on a significant scale, meaning the cells far away are responding systemically.

Ines: It's about how each cell type in that subcutaneous fascia served a different role in the deep learning framework.

Marcus: They identified a pathological macrophage-associated microenvironmental homeostasis alongside pericytes and other fibroblasts.

Yuki: What they found was impaired aerobic respiration in those macrophages, and they linked this to potential crosstalk with CPTCs during dysentery.

Ines: So the intercellular communication network inside the subcutaneous fascia showed an upregulation of macrophage proliferation mediated by CPTC-derived signaling.

Marcus: They pinpointed prominent ligand-receptor interactions like Wnt2/Wnt5a-Lrp6 and Fgf10/Fgfr1 as key drivers here.

Yuki: The macrophages in the DG showed augmented proliferative potential and a Warburg metabolic reprogramming, which mirrors what you see during enteric bacterial infections.

Ines: Specifically, they saw significant downregulation of the tricarboxylic acid cycle and an upregulation of gluconeogenesis pathways under dysentery conditions compared to the healthy control group.

Marcus: That change in cell cycle distribution was also notable; they observed a marked reduction in cells in the G1 phase and a concurrent increase in the S phase during infection.

Yuki: The regulatory mechanisms uncovered by deep learning pointed to three transcription factors that were significantly upregulated downstream of Wnt/Fgf signaling: Fosl1, Nfkb2, and Atf4.

Ines: GSEA functional enrichment showed those target genes were significantly enriched for functions related to cellular proliferation and DNA replication in the DG.

Marcus: And functions associated with aerobic respiration were markedly downregulated in that context.

Yuki: The diagnostic potential is strong because the transcriptional profile of subcutaneous fascia cells exhibits exceptional generalization performance when applied to the opposing group’s dataset.

Ines: Specifically, their G1 model achieved an accuracy of ninety-six point four four percent on the G2 test dataset, which really underscores how robust this biomarker is for visceral diseases.

Marcus: This suggests that subcutaneous fascia mediates the correlation between superficial tissues and visceral organs as a sensitive indicator of systemic homeostasis under various pathological conditions.

Yuki: The key finding on the FVICP is that it's underpinned by the systemic spread of signals from visceral organs and a synchronization of immune states across different tissues.

Ines: So we're seeing pathological equilibrium established in those remote local immune microenvironments in response to visceral dysregulation.

Marcus: They used Cell Ranger version seven point two.zero to generate the expression matrix and Seurat v4 point 4 for primary processing within the R environment version four point three.zero for this study on Salmonella dysentery induction in six female Sprague-Dawley rats receiving a dose of five percent sodium bicarbonate before infection, which was adjusted to five times ten eight colony-forming units per milliliter using zero point zero one M PBS.

Yuki: They also used Transmission Electron Microscopy to investigate functional interactions between macrophages and CPTCs, showing morphological evidence of those interactions.