Daily Summary for 2026-10-26

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

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

Marcus: 10 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 Shorter but not more catalytic.

The day's research: Ines: Today we focus on tau reshaping amyloid beta secondary nucleation, directly relating to plaque formation processes.

Marcus: AKMPC couples tau and amyloid beta into a model tracking fibril number and mass.

Yuki: Shorter tau-seeded fibrils do not inherently increase catalytic activity; the effect depends on site density or exposure.

Ines: That connects to the findings in From Turing Patterns to Spike Self-Replication and Nucleation in a Growing-Domain Tulip Breaking Virus Model.

Marcus: That work shows different mechanisms like spike self-replication versus nucleation depend on domain size.

Yuki: Synthetic data generation methods for single-cell RNA sequencing are being benchmarked against other models for privacy and biological fidelity.

Ines: Specifically, scGFT and scLDM are compared against others in Benchmarking Recent Single-Cell Synthetic Data Generation Methods on the OneK1K Dataset.

Marcus: Choices in handling non-highly variable genes are crucial for maintaining privacy while preserving biological structure there.

Yuki: Report the tolerance, not the count: exact phylogenetic distance deconvolution with incomplete references investigates how missing references affect inferring mixture-to-reference distances.

Ines: That investigation reveals which specific references are missing matters more than just the total number lost.

Ines: Music Between Two Minds suggests creative interaction needs a measurable organization.

Marcus: That implies individual voices share a common rhythmic state.

Yuki: How does that differ from simulation-based inference findings?

Ines: Gate-Based Feature Importance in Simulation-Based Inference for EEG shows learned gates can contradict ablation studies.

Marcus: So, which input features actually drive model performance is unclear then.

Yuki: The most relevant work concerns how community structure dictates stability.

Ines: Understanding what keeps microbial groups from collapsing under stress provides a blueprint for resilience in complex biological systems.

Marcus: The generalized modeling approach lets us test system stability across many models with varying sizes and structures.

Yuki: Community stability is constrained by self-limitation and metabolic efficiency factors.

Ines: This means hierarchical trophic structures actively counteract destabilizing cross-feeding loops, as seen in AKMPC.

Ines: This structural mediation shows organization stabilizes against perturbation because it is a key mechanism.

Marcus: The neuro-metabolic entanglement theory suggests mental experience is intertwined with physical-chemical constraints of the neural medium itself.

Yuki: That entanglement involves a recursive loop where forebrain processing influences midbrain metabolism, creating a bottleneck for unitary experience.

Ines: The computational models push beyond fixed energy supply; activity dynamically modulates its own metabolism through this proposed loop.

Marcus: This dynamic modulation means information processing and resource supply are recursively linked within the system.

Yuki: This concept offers a pathway to the explanatory gap by recognizing entanglement between neural organization and metabolic constraints.

Ines: Shorter but not more catalytic is worth hearing because it models how tau reshapes amyloid-beta secondary nucleation, showing cross-seeding accelerates fibrillation only when sites are numerous or exposed enough to be catalytic.

Marcus: Benchmarking Recent Single-Cell Synthetic Data Generation Methods on the OneK1K Dataset is valuable for comparing methods balancing biological accuracy and privacy protection.

Yuki: Report the tolerance, not the count is important because it shows missing specific references matter more than just how many are absent in phylogenetic analysis.

Ines: Music Between Two Minds: Individual Signatures, Shared States, and the Emergence of Creative Interaction is worth hearing as it finds creative interaction requires balancing unique musical voices with a shared rhythmic state.

Marcus: Modelling Toxin-Induced Intraspecific Personality Variation in Predator-Prey Dynamics shows how environmental toxins alter population stability through changes in predator and prey personalities.

Yuki: Gate-Based Feature Importance in Simulation-Based Inference for EEG is cautionary because learned feature importance weights from simulation models might not reflect true causal importance.

Ines: From Turing Patterns to Spike Self-Replication and Nucleation in a Growing-Domain Tulip Breaking Virus Model investigates whether viral patterns form by self-replication or by new spike nucleation as domain size grows.

Marcus: Dynamics of random recurrent neural networks: colored noise and sublinear activations examines how internal noise in recurrent neural networks affects their transition from local stability to chaos.

Yuki: Generalized modeling reveals stabilizing mechanisms in cross-feeding microbial communities because community stability is determined by resource interactions, self-limitation, and hierarchical structure.

Ines: The medium is the message: How neurometabolic entanglement generates conscious experience proposes a new theory suggesting consciousness arises from the link between neural information processing and brain metabolic energy.

Yuki: This paper is worth hearing as it suggests consciousness stems from the inseparable link between neural information processing and metabolic energy.

Ines: Indeed, that framework connects organization, metabolism, and subjective experience through recursive loops.

Lucky paper: 2610.30531: Ines: So we’re looking at Shorter but not more catalytic, which is an active-site-resolved model of how tau reshapes amyloid beta secondary nucleation.

Marcus: The central conclusion there is that tau can strongly increase oligomer-generating pressure without the tau-seeded fibrils themselves being intrinsically more active catalysts.

Yuki: But the effect isn't uniform; it totally depends on the local kinetic environment, specifically how much free monomer concentration the surrounding system controls.

Ines: That leads us right into that Damköhler number comparison between monomer consumption via elongation and monomer clearance in Shorter but not more catalytic.

Marcus: When Elongation Dominates, which is when the Damköhler number is one, tau coupling suppresses homotypic secondary nucleation even if the direct tau-catalyzed nucleation might initially raise oligomer-equivalent exposure.

Yuki: So it’s a switch based on whether elongation or clearance controls the monomer pool.

Ines: Exactly; and then there's that composite scenario where these two regimes occur successively as deposition progresses in that Shorter but not more catalytic model.

Marcus: The paper identifies four distinct routes through which tau can enhance oligomer generation pressure, including mass partitioning and site exposure.

Yuki: Mass partitioning is interesting because it suggests tau coupling can increase the total fibril mass available while keeping that clearance-controlled free monomer concentration.

Ines: That contrasts with the idea that shorter fibrils automatically mean more active sites per unit mass; the model shows you need a matched-mass comparison to test that.

Marcus: The authors are clear that they need a matched-mass comparison of control versus tau-seeded A beta forty-two fibrils to measure active-site density and site accessibility, not just length.

Yuki: They also mention dynamic site counting during different phases, which is necessary to distinguish the yields of birth sites versus growth sites in Shorter but not more catalytic.

Ines: That kinetic characterization using soluble A beta clearance and elongation rates is what they need to accurately determine the Damköhler number.

Marcus: I'm interested in how this context-dependent modulation of aggregation pathways actually changes what we think about plaque formation in real tissue.

Yuki: It means tau isn't just a simple catalyst enhancer; it’s a kinetic modulator that depends entirely on whether the system is clearing monomer or growing fibrils.

Ines: That puts things into perspective for how we model these complex biological systems, moving beyond just counting fibril length to understanding the underlying chemistry of site accessibility.

Lucky paper: 2610.31454: Ines: So we’re digging into Benchmarking Recent Single-Cell Synthetic Data Generation Methods on the OneK1K Dataset today, which looks at scGFT, scLDM, and Cell2Sentence.

Marcus: This paper is looking at the trade-off between biological utility and privacy risk when creating synthetic single-cell RNA sequencing data.

Yuki: They restricted the modeling phase to highly variable genes to capture the main axes of biological variation in that OneK1K dataset, which has about one point two million peripheral blood mononuclear cells from nine hundred eighty-two donors.

Ines: scLDM is a two-stage latent-diffusion generator where a transformer VAE compresses expression into a low-dimensional continuous latent space, and then a Diffusion Transformer denoises that under classifier-free guidance on cell type.

Marcus: But they trained scLDM only on the top one thousand highly variable genes to get that output, which means the raw synthetic data doesn't align with the full OneK1K test set.

Yuki: They tested two variants for scLDM: the raw HVG-only output and a hybrid that stacks those counts with Poisson samples for the remaining twenty-four thousand eight hundred thirty-four non-HVG genes.

Ines: The scGFT framework is training-free and uses the Discrete Fourier Transform to map gene expression profiles into a complex space, treating it like a graph signal.

Marcus: The hyperparameter controlling how many components are modified in scGFT governs the trade-off between diversity and biological fidelity; higher values explore more diverse states.

Yuki: For privacy, they made a modified version of scGFT where non-HVG gene expression is synthesized by sampling from a Poisson distribution parameterized by mean expression levels estimated from the training data.

Ines: The results show that the Poisson baseline achieved the highest Spearman correlation for scGFT comparison at zero point nine nine eight.

Marcus: However, both scGFT and scLDM better preserve the underlying biological manifold in downstream evaluations, as shown by their GT vs. Comb ARI metrics.

Yuki: Specifically, scGFT achieves an ARI of zero point three four two compared to the Poisson baseline's zero point two two zero in that evaluation metric.

Ines: Regarding privacy risk, for scGFT, copying non-HVG variants is non-viable, reaching a perfect membership inference leakage AUC of one point zero because of the method they used.

Marcus: Increasing the number of Fourier components to n = twenty-five in scGFT reduces that leakage to a near-random privacy profile with an AUC of zero point four nine four, which is comparable to the Poisson baseline.

Yuki: Similarly, scLDM also attains an AUC of zero point five zero zero in the privacy evaluation, indicating a random distribution and demonstrating both models meet the privacy constraints for donor-level risk.

Ines: The study concludes that both scGFT and scLDM substantially exceed the Poisson baseline in capturing joint structure that independent sampling cannot achieve.

Marcus: This motivates a broader evaluation perspective for Track II to prioritize robustness across multiple metrics, reporting both HVG-focused and full-gene behavior.

Lucky paper: 2610.30776: Ines: We’re looking at Report the tolerance, not the count, which is paper number three of five this show.

Marcus: It focuses on how many strictly interior sources are required to explain a phylogenetic distance profile when some references are missing.

Yuki: The gist is that the minimum number of strictly interior sources needed is more informative than just providing an exact count, because it shows how much tolerance there is for errors in reference data.

Ines: The model defines a distance from the whole mixture to each reference as a weighted average of source-to-reference tree distances.

Marcus: The problem asks if a given distance profile admits an explanation with at most K sources and if so, what the smallest such explanation is.

Yuki: It involves exact rational data where sources sit at points in the interiors of original edges with positive weights and a positive mean pendant.

Ines: The paper provides an exact algorithm for a free positive mean pendant and any nonempty set of observed leaves on a positive-rational reduced tree, achieving two O(K) poly(B) bit complexity.

Marcus: But with arbitrary missing references and a prescribed pendant, the problem becomes Wone-hard in the source budget K, even on strictly binary trees supplied with a strict explanation using at most K plus one sources.

Yuki: The complexity classification distinguishes between different information regimes; for instance, any nonempty subset Free y greater than zero has a deterministic algorithm using O(N) plus 2O(K) rational arithmetic operations and two O(K) poly(B) bit time.

Ines: The same bounds hold with a prescribed positive pendant when at most one original leaf is unobserved.

Marcus: All, or all but one Prescribed y0 greater than zero has the same upper bound as the free-positive-pendant problem, being NP-complete and no subexponential K dependence under ETH.

Yuki: Arbitrary subset Prescribed y0 = one is Wone-hard in K, even with a supplied strict (K plus one)-source explanation, and the lower bound matches the linear order of the exponent in K, not an optimal exponential base.

Ines: They analyze this geometry using vertex potentials and a baseline definition; they define the lower Lipschitz envelope h(v) and baseline bY which are used to bound the deficit term.

Marcus: The lemma states that if at most K sources realize f, then S is empty for K = one and also for K = two when every nonleaf vertex has degree at least three.

Yuki: For K greater than or equal to two the assertion is that X v in S (degree of v minus two) less than or equal to K minus two.

Ines: They investigate how removing references affects the minimum source count, finding that "the direction is a theorem at every error tolerance and held without exception in twenty-three thousand five hundred forty-five draws."

Marcus: Two adversarially chosen references suffice, at the median, to destroy uniqueness of the minimum explanation, and that number does not grow with panel size.

Yuki: Discarding seven thousand nine hundred thirty-one of ten thousand five hundred seventy-five references from a real 16S panel at random costs one point eight percent of the tolerance at which one source suffices.

Ines: The phenomenon is not unobservable in practice; the exact minimum fell in four hundred fifty of one thousand nine hundred sixty-eight draws on one hundred one real reference trees, by up to six sources.

Marcus: The work establishes neither bounded-aspect-ratio hardness nor robustness to observational error.

Yuki: The final conclusion is that the tolerance at which an explanation of a given size exists is what should be reported, based on exact rational data and a known metric tree.

Lucky paper: 2610.31131: Ines: So we’re looking at Music Between Two Minds, which is paper number four of five this day, exploring how two minds jointly regulate the flow of novelty during improvisation.

Marcus: The core idea here is that improvising pairs don't fuse into one system or stay independent; they preserve individual melodic voices while sharing a rhythmic state.

Yuki: The study used nine hundred eighty-six professional jazz improvisations to see how two minds jointly regulate the flow of novelty in real time.

Ines: They used a general predictive model of jazz expectation, trained on the Weimar Jazz Database, to measure pitch-based and rhythm-based fluctuations for every single note.

Marcus: Pitch-based fluctuations behaved as stable signatures of individual identity, meaning "the two partners differed within ninety percent of takes" and "performer identity could be decoded from single solos."

Yuki: But then you have the rhythm-based fluctuations, which showed the reverse organization, where "partners were statistically indistinguishable within a take yet shifted together from take to take."

Ines: So the shared rhythmic state isn't static; it drifts gradually across each performance as both performers move together.

Marcus: That drift is evidenced by tracking how dissimilarity between solos grew with the number of turns separating them, showing that "both series climbed together" in rhythm.

Yuki: And at every solo hand-off, they showed a systematic push–pull complementarity where "when one performer exceeded their own average surprise in the predictive dynamics of improvisation, the partner's next solo fell below their own."

Ines: That coupling was found to be reliable because real duos showed "reliable negative coupling that both controls abolished" when compared against null models.

Marcus: The central pattern is a double dissociation where "melodic fluctuations carried a persistent individual identity: partners never converged, their separation neither narrowed with turns nor shifted with takes."

Yuki: While the rhythm-based fluctuations carried a collective state, which was "indistinguishable between partners, specific to each real interaction."

Ines: This organization is described as field intelligence, which is "structure in predictive fluctuations that cannot be attributed to either partner alone."

Marcus: How does that division of roles across dimensions allow a duo to be "maximally close and reliably distinct at the same time"?

Yuki: The paper concludes that interpersonal improvisation depends not on convergence toward a single common state, but on the simultaneous maintenance of individual and shared predictive dynamics.

Ines: The limitations include the fact that "the melodic identity signature may partly reflect instrument-specific affordances rather than personal style alone."

Marcus: And they also note that "all performances were improvisations on a single tune within a single idiom," which means the findings might not hold across all repertoires or genres.

Yuki: The analysis is restricted to symbolic note sequences, meaning the findings may not hold across all interactions that can be represented as a sequence of events under a shared predictive model.

Ines: They also state that the reference model was trained on a specific historical corpus of monophonic solos, and different reference corpora could shift absolute values of surprise and uncertainty.

Marcus: The results don't yet show that this regulation is what makes the interaction creative in the evaluative sense, just that novelty is jointly regulated.

Yuki: They suggest that "some informational dimensions of speech should identify the speaker and resist convergence, others should belong to the particular conversation and drift with it."

Ines: The findings show how novelty is jointly regulated, but they do not yet show that this regulation is what makes the interaction creative in the evaluative sense.

Lucky paper: 2610.30431: Ines: So we're looking at Modelling Toxin-Induced Intraspecific Personality Variation in Predator-Prey Dynamics, which explores how personality traits can qualitatively change predator and prey interactions when exposed to environmental toxins like methylmercury.

Marcus: The study extends a toxin-dependent model by incorporating personality variation in both species; the prey's personality affects its birth and growth rate, while the predator's personality changes its functional response.

Yuki: It parameters this for rainbow trout and aquatic insect prey exposed to methylmercury, looking at biomass densities x for prey and y for predator, along with toxin concentrations U and V.

Ines: The paper shows that prey personality tends to reduce biomass fluctuations and promote stable coexistence, which is interesting because we saw that in other models.

Marcus: In contrast, predator personality can increase biomass fluctuations and lead to persistent population oscillations at low and intermediate toxin concentrations.

Yuki: When looking at the mathematical analysis for the paper Modelling Toxin-Induced Intraspecific Personality Variation in Predator-Prey Dynamics, it reduces the system to two dimensions using a quasi-steady-state approximation.

Ines: That reduction establishes the positivity and boundedness of the reduced system and identifies a positively invariant region, which is a key detail.

Marcus: The equilibrium structure depends jointly on the toxin level T and the prey and predator personality contrasts, specifically alpha for prey contrast and rho for predator contrast.

Yuki: The relative personality contrast for predators is defined as rho = mu1 - mu2 / two where is the average of those two personalities.

Ines: And the relative prey personality contrast is defined as alpha = alpha11 - alpha12 / two which we see directly impacts stability.

Marcus: Increasing the prey personality contrast, alpha, decreases oscillation amplitudes and promotes stability, while increasing the predator contrast, rho, destabilizes the system by generating and amplifying oscillations.

Yuki: This suggests that differences in birth and growth in prey suppress oscillations, but differences in predator foraging generate them.

Ines: The paper Modelling Toxin-Induced Intraspecific Personality Variation in Predator-Prey Dynamics concludes that toxin-induced personality variation can fundamentally change long-term dynamics before direct toxic effects cause population loss.

Marcus: That finding is significant because it suggests that behavioral traits can alter population dynamics even at low and moderate toxin concentrations relevant to freshwater ecosystems.