Benjamin B. Machta

Correlation functions quantify super-resolution images and estimate apparent clustering due to over-counting.

We present an analytical method using correlation functions to quantify clustering in super-resolution fluorescence localization images and electron microscopy images of static surfaces in two dimensions. We use this method to quantify how over-counting of labeled molecules contributes to apparent self-clustering and to calculate the effective lateral resolution of an image. This treatment applies to distributions of proteins and lipids in cell membranes, where there is significant interest in using electron microscopy and super-resolution fluorescence localization techniques to probe membrane heterogeneity. When images are quantified using pair auto-correlation functions, the magnitude of apparent clustering arising from over-counting varies inversely with the surface density of labeled molecules and does not depend on the number of times an average molecule is counted. In contrast, we demonstrate that over-counting does not give rise to apparent co-clustering in double label experiments when pair cross-correlation functions are measured. We apply our analytical method to quantify the distribution of the IgE receptor (FcεRI) on the plasma membranes of chemically fixed RBL-2H3 mast cells from images acquired using stochastic optical reconstruction microscopy (STORM/dSTORM) and scanning electron microscopy (SEM). We find that apparent clustering of FcεRI-bound IgE is dominated by over-counting labels on individual complexes when IgE is directly conjugated to organic fluorophores. We verify this observation by measuring pair cross-correlation functions between two distinguishably labeled pools of IgE-FcεRI on the cell surface using both imaging methods. After correcting for over-counting, we observe weak but significant self-clustering of IgE-FcεRI in fluorescence localization measurements, and no residual self-clustering as detected with SEM. We also apply this method to quantify IgE-FcεRI redistribution after deliberate clustering by crosslinking with two distinct trivalent ligands of defined architectures, and we evaluate contributions from both over-counting of labels and redistribution of proteins.

Parameter Space Compression Underlies Emergent Theories and Predictive Models

Information Physics Multiparameter models, which can emerge in biology and other disciplines, are often sensitive to only a small number of parameters and robust to changes in the rest; approaches from information theory can be used to distinguish between the two parameter groups. In physics, on the other hand, one does not need to know the details at smaller length and time scales in order to understand the behavior on large scales. This hierarchy has been recognized for a long time and formalized within the renormalization group (RG) approach. Machta et al. (p. 604) explored the connection between two scales by using an information-theoretical approach based on the Fisher Information Matrix to analyze two commonly used physics models—diffusion in one dimension and the Ising model of magnetism—as the time and length scales, respectively, were progressively coarsened. The expected “stiff” parameters emerged, in agreement with RG intuition. An information-theoretical approach is used to distinguish the important parameters in two archetypical physics models. The microscopically complicated real world exhibits behavior that often yields to simple yet quantitatively accurate descriptions. Predictions are possible despite large uncertainties in microscopic parameters, both in physics and in multiparameter models in other areas of science. We connect the two by analyzing parameter sensitivities in a prototypical continuum theory (diffusion) and at a self-similar critical point (the Ising model). We trace the emergence of an effective theory for long-scale observables to a compression of the parameter space quantified by the eigenvalues of the Fisher Information Matrix. A similar compression appears ubiquitously in models taken from diverse areas of science, suggesting that the parameter space structure underlying effective continuum and universal theories in physics also permits predictive modeling more generally.

Why are Nonlinear Fits to Data so Challenging

Fitting model parameters to experimental data is a common yet often challenging task, especially if the model contains many parameters. Typically, algorithms get lost in regions of parameter space in which the model is unresponsive to changes in parameters, and one is left to make adjustments by hand. We explain this difficulty by interpreting the fitting process as a generalized interpolation procedure. By considering the manifold of all model predictions in data space, we find that cross sections have a hierarchy of widths and are typically very narrow. Algorithms become stuck as they move near the boundaries. We observe that the model manifold, in addition to being tightly bounded, has low extrinsic curvature, leading to the use of geodesics in the fitting process. We improve the convergence of the Levenberg-Marquardt algorithm by adding geodesic acceleration to the usual step.

Liquid General Anesthetics Lower Critical Temperatures in Plasma Membrane Vesicles

A large and diverse array of small hydrophobic molecules induce general anesthesia. Their efficacy as anesthetics has been shown to correlate both with their affinity for a hydrophobic environment and with their potency in inhibiting certain ligand-gated ion channels. In this study we explore the effects that n-alcohols and other liquid anesthetics have on the two-dimensional miscibility critical point observed in cell-derived giant plasma membrane vesicles (GPMVs). We show that anesthetics depress the critical temperature (Tc) of these GPMVs without strongly altering the ratio of the two liquid phases found below Tc. The magnitude of this affect is consistent across n-alcohols when their concentration is rescaled by the median anesthetic concentration (AC50) for tadpole anesthesia, but not when plotted against the overall concentration in solution. At AC50 we see a 4°C downward shift in Tc, much larger than is typically seen in the main chain transition at these anesthetic concentrations. GPMV miscibility critical temperatures are also lowered to a similar extent by propofol, phenylethanol, and isopropanol when added at anesthetic concentrations, but not by tetradecanol or 2,6 diterbutylphenol, two structural analogs of general anesthetics that are hydrophobic but have no anesthetic potency. We propose that liquid general anesthetics provide an experimental tool for lowering critical temperatures in plasma membranes of intact cells, which we predict will reduce lipid-mediated heterogeneity in a way that is complimentary to increasing or decreasing cholesterol. Also, several possible implications of our results are discussed in the context of current models of anesthetic action on ligand-gated ion channels.

Critical casimir forces in cellular membranes

Recent experiments suggest that membranes of living cells are tuned close to a miscibility critical point in the two-dimensional Ising universality class. We propose that one role for this proximity to criticality in live cells is to provide a conduit for relatively long-range critical Casimir forces. Using techniques from conformal field theory we calculate potentials of mean force between membrane bound inclusions mediated by their local interactions with the composition order parameter. We verify these calculations using Monte Carlo simulations where we also compare critical and off-critical results. Our findings suggest that membrane bound proteins experience weak yet long-range forces mediated by critical composition fluctuations in the plasma membranes of living cells.

Experimental Observations of Dynamic Critical Phenomena in a Lipid Membrane

Near a critical point, the time scale of thermally induced fluctuations diverges in a manner determined by the dynamic universality class. Experiments have verified predicted three-dimensional dynamic critical exponents in many systems, but similar experiments in two dimensions have been lacking for the case of conserved order parameter. Here we analyze the time-dependent correlation functions of a quasi-two-dimensional lipid bilayer in water to show that its critical dynamics agree with a recently predicted universality class. In particular, the effective dynamic exponent z(eff) crosses over from ~2 to ~3 as the correlation length of fluctuations exceeds a hydrodynamic length set by the membrane and bulk viscosities.

Programming a Parallel Computer: The Ersatz Brain Project

There is a complex relationship between the architecture of a computer, the software it needs to run, and the tasks it performs. The most difficult aspect of building a brain-like computer may not be in its construction, but in its use: How can it be programmed? What can it do well? What does it do poorly? In the history of computers, software development has proved far more difficult and far slower than straightforward hardware development. There is no reason to expect a brain like computer to be any different. This chapter speculates about its basic design, provides examples of “programming” and suggests how intermediate level structures could arise in a sparsely connected massively parallel, brain like computer using sparse data representations.

Parallel Dynamics and Computational Complexity of Network Growth Models

The parallel computational complexity or depth of growing network models is investigated. The networks considered are generated by preferential attachment rules where the probability of attaching a new node to an existing node is given by a power alpha of the connectivity of the existing node. Algorithms for generating growing networks very quickly in parallel are described and studied. The sublinear and superlinear cases require distinct algorithms. As a result, there is a discontinuous transition in the parallel complexity of sampling these networks corresponding to the discontinuous structural transition at alpha=1 , where the networks become scale-free. For alpha>1 , networks can be generated in constant time while for 0</=alpha<1 , logarithmic parallel time is required. The results show that these networks have little depth and embody very little history dependence despite being defined by sequential growth rules.

Power-law velocity distributions in granular gases.

The kinetic theory of granular gases is studied for spatially homogeneous systems. At large velocities, the equation governing the velocity distribution becomes linear, and it admits stationary solutions with a power-law tail, f (v) approximately v(-sigma) . This behavior holds in arbitrary dimension for arbitrary collision rates including both hard spheres and Maxwell molecules. Numerical simulations show that driven steady states with the same power-law tail can be realized by injecting energy into the system at very high energies. In one dimension, we also obtain self-similar time-dependent solutions where the velocities collapse to zero. At small velocities there is a steady state and a power-law tail but at large velocities, the behavior is time dependent with a stretched exponential decay.

Dissipation Bound for Thermodynamic Control.

Biological and engineered systems operate by coupling function to the transfer of heat and/or particles down a thermal or chemical gradient. In idealized deterministically driven systems, thermodynamic control can be exerted reversibly, with no entropy production, as long as the rate of the protocol is made slow compared to the equilibration time of the system. Here we consider fully realizable, entropically driven systems where the control parameters themselves obey rules that are reversible and that acquire directionality in time solely through dissipation. We show that when such a system moves in a directed way through thermodynamic space, it must produce entropy that is on average larger than its generalized displacement as measured by the Fisher information metric. This distance measure is subextensive but cannot be made small by slowing the rate of the protocol.