Purushottam Dixit

Big picture: Directional sensing can be a receptor-level computation driven by diffusion + degradation + feedback.
This explains puzzling experiments (e.g., why blocking endocytosis impairs chemotaxis) and suggests a general biophysical strategy across receptors families.

The stochastic analysis is especially striking:
Noise reveals an optimal basal activity set point that maximizes signal-to-noise. Too little activity → noisy. Too much → less directional contrast. Biology tunes itself in between.

Predictions:
✔ There is an optimal diffusion rate for maximal polarization
✔ Receptors encode relative, not absolute, ligand gradients
✔ Cooperative receptor interactions can amplify weak gradients
All without requiring separate “local” and “global” signaling species.

Add diffusion, and something remarkable happens:
Receptors move from low-ligand regions to high-ligand regions, but degradation depletes them where ligand is high.
This mismatch creates polarized receptor activity-directional sensing emerges naturally.

Without diffusion, degradation implements integral feedback: receptor activity adapts perfectly to a ligand-independent set point-essential for sensing relative changes rather than absolute ligand levels.

The key ingredients:
• Receptor diffusion along the membrane
• Basal (ligand-independent) activity
• Selective degradation of active receptors
Together, these create an integral feedback loop (as opposed to an IFFL in LEGI) at the cell surface.

Classic models of eukaryotic directional sensing (e.g., LEGI) assume receptors are passive: they report ligand levels, while intracellular networks compare “local vs global” signals using incoherent feedforward loops. We asked: can receptors do the computation directly?

How do cells know which way to move in a chemical gradient? 🧭 New work by graduate student Andrew Goetz proposes that receptors can compute direction. This new mechanism for directional sensing was published in PNAS late last year: www.pnas.org/doi/10.1073/...

Given the centralization of information flow, our analysis suggests that re-tuning just a few TFs could in principle rejuvenate information flow and restore gene expression. Aging may not be just cellular damage, but a gradual communication breakdown.

While signs of cellular aging may be varied, across all tissues, mutual information between TFs and TGs declines with age. Aged networks show input mismatch, higher centralization, and reduced stability, patterns reminiscent of aging brains and failing ecosystems.

We borrowed statistical physics/neuroscience models to study information transmission in noisy systems.
We treated transcription factors (TFs) → target genes (TGs) as a multi-input, multi-output communication channel, and measured its fidelity using mutual information.

We analyzed single-cell RNA-seq data from multiple mouse tissues across the lifespan. This dataset captures how thousands of genes are expressed in individual cells, letting us see how regulatory communication changes with age.

In our new paper led by Brooke Emison, in collaboration with Fabrisia Ambrosio, Andrew Mugler, and @chriswlynn.bsky.social, we show that as cells age, the flow of transcriptional information in gene regulatory networks breaks down.

What if aging isn’t just cellular damage, but a lossy transmission of information?
We used single-cell data and physics-based models to show that as cells age, the flow of transcriptional information collapses. Here’s what that means. 👇

We look forward to submitting our revision. As always, our experience with the new reviewing model of @elife.bsky.social has been wonderful! 6/n n =6

The implications?
It’s not just who binds tighter, but who survives the network's non-equilibrium processing.

Ligand-specificity is an emergent property of the entire network architecture not just binding thermodynamics 5/n

This behavior emerges only when the system is driven out of equilibrium. Energy dissipation (via dissociation and degradation) enables sharp ligand discrimination—not possible in equilibrium systems. 4/n

This means:
➡️Increasing ligand affinity can decrease signaling.
➡️The system has an optimal “sweet spot” for specificity and kinase activity
➡️Ligands with similar affinities can produce very different outputs depending on cellular parameters 3/n

We built a minimal model of receptor signaling that includes common signaling receptor features: Multi-site phosphorylation, rapid dissociation, and Ligand-dependent receptor degradation. Together, they create non-monotonic responses to ligand affinity and kinase activity. 2/n

It’s often assumed that stronger ligand binding = stronger signaling with non-equilibrium effects further enhancing this preference (a.k.a. kinetic proofreading). But often, thermodynamics preference is reversed! We asked: could non-equilibrium mechanisms help explain why? 1/n

Just out: new preprint (with @ralitsamadsen.bsky.social) is now #Reviewed at @eLife! "Non-equilibrium strategies for ligand specificity in signaling networks". We show how cells use non-equilibrium strategies to discriminate between ligands in surprising ways 🔗 elifesciences.org/reviewed-pre...

We have a funding for postdoctoral fellowship that needs to be filled very soon. We are exploring new directions (1) at the interface of non-equilibrium computations and physical learning and (2) in building ecologically motivated machine learning models for microbiomes. Please spread the word!

Accessible surface area - Wikipedia

What about accessible surface area? You get to choose the size of the "probe" which may be useful: en.wikipedia.org/wiki/Accessi...

Just learnt that our collaborative grant proposal on identifying master regulators of T cell metabolism in Lupus will not be reviewed as the study section was indefinitely postponed, along with many other grants that are vital to biomedical research.

Ugh that sucks! Let's hope for the best..

Thank you! We submitted one last December, so it technically is before the expiration date in Jan 2025.

Has anybody looked at approximate Bayesian computation (ABC) using stat mech? The formulation used in rejection ABC: rho(S(D),S(D')) < eps (rho is discrepancy function, S is a summary stat, D is data, D' is simulated data) looks an awful lot like the microcanonical ensemble with an energy = S(D).

Designing host-associated microbiomes using the consumer/resource model | mSystems Generative models are extremely popular in modern biology. They have been used to model the variation of protein sequences, entire genomes, and RNA sequencing profiles. Importantly, generative models ...

Our work on computational design of microbiomes with desired host properties using mechanistic models is now out in mSystems: journals.asm.org/doi/10.1128/...

Analysis of available signaling network parameters suggests that LAGS is widely applicable. Moreover, preferential degradation is just one mechanism for integral feedback control. Therefore, other habituation mechanisms e.g. activity induced inactivation, should also work the same way!

Additionally, when combined with receptor oligomerization, an increase in preferential degradation allows cells to sense relative ligand gradients over a larger range of background ligand concentrations. This is sometimes known as the Weber-Fechner law.