Time | Speaker | Title | Resources | |
---|---|---|---|---|
13:30 to 13:45 | -- | Introductory session | ||
13:45 to 14:00 | Recorded talk (APS) |
(MLH) Self-organization of microtubules and motors drive large-scale intracellular flows in cells Presenter: Reza Farhadifar (Simons Foundation) Cytoplasmic streaming is essential for transporting and mixing nutrients, proteins, and organelles within large plant and animal cells. The large ~200um Drosophila oocyte has recently gained attention for experimental and theoretical studies of this phenomenon. We present a quantitative study of streaming in Drosophila oocytes that combines PI of 3D time-lapse movies, with biophysical modeling and simulation. We observe a diverse family of 3D vortical flows across different oocytes, which differ in position and orientation, and which last tens of minutes. We show that a model of cytoskeletal activity at the periphery, organized by its interaction with interior fluid, explains the observed streaming structures. The emerging picture sheds light on a class of intracellular flows in large cells and highlights the wealth of questions at the interface of geometry, active matter, and basic biology. |
||
13:45 to 14:00 | Ishita Shitut (IISc, India) |
(RLH) Real-Space Stochastic GW Calculations Benchmark on GW100 Stochastic implementation of GW is a linear scaling method, ideally suited for calculating quasi-particle energies of large systems. This approach uses the stochastic resolution of identity to represent Green’s function as a product of a randomly generated orbital at time zero and an evolved random orbital at a later time. It employs real time propagation of stochastic functions to obtain screened coulomb response function. The response function is efficiently stored using stochastic compression. We have implemented the stochastic GW method in real-space density functional theory code PARSEC. We have benchmarked our stochastic GW implementation on GW100 set against the results obtained from the NanoGW code [1]. We find that our results are in good agreement with the results obtained from the NanoGW code. [1] W. Gao, and J. Chelikowsky, J. Chem. Theory Comput. 15, 5299 (2019) ACKNOWLEDGMENTS |
||
14:00 to 14:45 | Recorded talk (APS) |
(MLH) Geometry and Genetics The application of quantitative methods to biological problems faces the choice of how much detail to include and the generality of the conclusions. Both routine data analysis and airy pronouncements that have nothing to say about everything are to be avoided. The middle ground entails some use phenomenology, a well-used approach in both high and low energy physics. A sampling of examples will be presented from my work in the area of developmental biology, to give a flavor of what is possible. They include experiments in synthetic embryology where human stem cells are coaxed into making patterns and structures similar to real embyos, use of modern (ie post 1960) mathematics to enumerate categories of dynamical behaviors, and a bit of computational evolution to address the question of what pattern forming systems can be evolved by incremental changes. |
||
14:00 to 14:15 | Kartick Ramakrishnan (IISc, India) |
(RLH) Scalable real-space finite-element based DFT calculations: Application to energy materials Ab initio material modelling using density functional theory (DFT) has provided critical insights into possible new chemical structures, thermodynamic and atomic properties, enabling the design of efficient energy storage materials. Further, high throughput DFT calculations have been instrumental in understanding the physical mechanisms that control the capacity, performance, safety, and longevity of energy storage devices. However, understanding certain nanoscale phenomena in these devices, that are inaccessible by experimental investigations, demand large scale DFT calculations with arbitrary boundary involving thousands of atoms. For example -- understanding energetics of non-stochiometric charged electrolyte surfaces, modelling electrode||electrolyte interfaces in presence of external electric field, predicting bonding information in large nano particles for screening hydrogen storage materials etc, all require computationally efficient, accurate DFT calculations scalable on evolving heterogeneous computing architectures. This requires a departure from conventionally used plane-wave (PW) or atomic orbital (AO) based codes, that restrict the simulation domains to periodic boundary conditions (PW) or lack a systematically convergent behaviour of basis sets (AO) and do not take the full advantage of existing heterogeneous parallel computing architectures. To this end, the recently proposed finite-element based methodologies for density functional theory (DFT-FE) addresses the above shortcomings, thereby providing a generic scalable computational framework to address challenging material modelling problems encountered in the design of energy storage devices. In this talk, I will be first discussing few benchmarking studies of the existing DFT-FE framework to calculate formation energies of various battery electrolyte material. Furthermore, the implementation and the benchmarking of nudged elastic band, AIMD and projected finite- element orbital population analysisframework within the DFT-FE formalism will be discussed. Finally, I will conclude with outlining the plans for computational methodology development in the DFT-FE framework for addressing the aforementioned challenges, enabling the design of energy storage devices. |
||
14:15 to 14:30 | Abhishek Kumar Adak (JNCASR, India) |
(RLH) Insights from density functional theory into the formation and rotation of an enantiospecific assembly of molecular raffle wheels We have performed a joint theoretical and experimental study of the assembly formed when BPP-COOH ( 2,6-bis(1H-pyrazol-1-yl)pyridine-4-carboxylic acid) is deposited on Ag(111). Three-fourths of the molecules form a rigid Kagome 'host' network. The cavities of this network are occupied by the remaining 'guest' molecules, which display a punctuated rotation between positions corresponding to global minima in the rotational energy landscape. Calculations show that the topography of this landscape can be explained by the making and breaking of hydrogen bonds between the guest molecules and the host network. The height of the rotational barrier computed theoretically is in excellent agreement with that extracted from temperature-dependent experiments. The host network also bestows enantiospecificity on the system, due to the twist between the host network and the underlying Ag(111) surface. ref. https://doi.org/10.1002/ange.202107708 |
||
14:30 to 14:45 | Arka Bandyopadhyay (IISc, India) |
(RLH) Electrically switchable giant Berry curvature dipole in silicene, germanene and stanene The anomalous Hall effect in time-reversal symmetry broken systems is underpinned by the concept of Berry curvature in band theory. However, recent experiments [1] reveal that the nonlinear Hall effect can be observed in non-magnetic systems without applying an external magnetic field [2]. The emergence of fascinating nonlinear Hall effect under time-reversal symmetric conditions can be well-explained in terms of non-vanishing Berry curvature dipole, i.e., the dipole moment of Berry curvature arising from inversion symmetry breaking [3, 4]. In this work [5], we have systematically availed realistic tight-binding models and symmetry analyses for the quantitative estimation of Berry curvature dipole. It has been observed that the combined effect of transverse electric field and strain leads to a giant Berry curvature dipole in the elemental buckled honeycomb lattices – silicene, germanene, and stanene [6]. In particular, the external electric field breaks the inversion symmetry of these systems, while strain helps to attain an asymmetrical distribution of Berry curvature of a single valley. Furthermore, the topology of the electronic wavefunction switches from the band inverted quantum spin Hall state to the normal insulating one at the gapless point. We have explored that this band gap closing at the critical electric field strength is accompanied by an enhanced Berry curvature and concomitantly a giant Berry curvature dipole at the Fermi level. Our results predict the occurrence of an electrically switchable nonlinear Hall effect in a new class of elemental systems that can be experimentally verified. |
||
14:45 to 15:00 | Recorded talk (APS) |
(MLH) Sensing and making sense of fluctuating cellular states The self-organisation of cells into complex tissue relies on the tight regulation of cellular behavior. Typically, the regulation of cell decisions is attributed to pathways controlling the concentration of molecular species in response to intrinsic or extrinsic signals, such as in gene regulatory networks. Here, by contrast, we show in the paradigmatic example of cell death that cells manipulate how fluctuations propagate across spatial scales to regulate cellular behavior. Specifically, we find that the feedback between molecular and mesoscopic organelle fluctuations gives rise to a quasi-particle degree of freedom whose intriguing kinetic properties construct a kinetic low-pass filter of time- dependent concentrations of signaling molecules. This allows cells to distinguish |
||
14:45 to 15:00 | Reshma Devi (IISc, India) |
(RLH) Effect of Exchange-Correlation Functionals on the Estimation of Migration Barriers in Battery Materials Facile ionic mobility within-host frameworks is crucial in designing high-energy-density batteries with high-power densities, where the migration barrier (Em) is the governing factor. This talk will discuss the accuracy and computational performance in calculated Em, against experimental data, of several exchange-correlation (XC) functionals, within the density functional theory-nudged elastic band framework of six different electrodes and three diverse solid electrolytes. The generalized gradient approximation (GGA), the strongly constrained and appropriately normed (SCAN), and their Hubbard U corrections, GGA+U and SCAN+U, are the important XC functionals considered. It is observed that SCAN tends to be more accurate than other frameworks, albeit with higher computational costs and convergence difficulties, while GGA is a feasible choice for ’quick’ and ’qualitative’ Em predictions. The sensitivity of Em on adding uniform background charge and/or the climbing image approximation in solid electrolytes and the Hubbard U correction in electrodes are also quantified. This benchmarking will thus aid in selecting the suitable XC functional for a given structure in future studies, thus enabling the discovery of novel ion-conducting electrodes and solid electrolytes via computational workflows. |
||
15:00 to 15:15 | Recorded talk (APS) |
(MLH) Mobile defects born from an energy cascade shape the locomotive behavior of a headless animal The physics of behavior seeks simple descriptions of animal behavior. The field has advanced rapidly by using techniques in low dimensional dynamics distilled from computer vision. Yet, we still do not generally understand the rules which shape these emergent behavioral manifolds in the face of complicated neuro-construction --- even in the simplest of animals. In this work, we introduce a non-neuromuscular model system which is complex enough to teach us something new but also simple enough for us to understand. We discover manifolds underlying the governing dynamics shaped and stabilized by a physical mechanism: an active-elastic, inverse-energy cascade. We explore the formulation of the governing dynamics of a polarized active elastic sheet in terms of the normal modes of an elastic structure decorated by a polarized activity at every node. By incorporating a torque mediated coupling physics, we show that the power is pumped from the shortest length scale up to longer length scale modes via a combination of direct mode coupling and preferential dissipation of higher frequency modes. We use this result to motivate the study of organismal locomotion as an emergent simplicity governing organism-scale behavior. To master the low dimensional dynamics on this manifold, we present a zero-transients limit study of the dynamics of +1 or vortex like defects in the ciliary field (which is experimentally supported for small organisms). We show, experimentally, numerically and analytically that these defects arise from this energy cascade to generate long-lived, stable modes of locomotive behavior. Using a geometric model, we show how the defect undergoes unbinding. We extend this framework as a tool for studying larger organisms with non-circular shape and introduce local activity modulation for defect steering. We expect this work to inform the foundations of organismal control of distributed actuation without muscles or neurons. |
||
15:00 to 15:15 | Nimish Dwarkanath (JNCASR, India) |
(RLH) Ab initio molecular dynamics: a key to unravel microscopic phenomena in metal-organic framework solids Density functional theory (DFT) based calculations are employed to study a variety of systems. In some cases, it is necessary to incorporate the finite temperature effects, and simple geometry optimizations, essentially 0 K calculations, do not suffice. I will introduce DFT-based ab initio molecular dynamics (AIMD) as a tool to study porous metal-organic framework (MOF) crystals with guest molecules based on our investigation. Summaries of the two research problems--(i) Recently reported halogenated MOFs1 showed the second step in CO2 adsorption isotherms at 195 K. Second isotherm step is typically exhibited by porous materials with large pores or those that swell upon adsorption resulting in significant lattice expansion; neither of them occurs in the halogenated MOF. The underlying microscopic cause for the second isotherm step was elucidated using AIMD simulations2. (ii) AIMD simulations revealed that the small pore channels of aluminium-based MOF (Al-NDC) spontaneously close, thus, guiding us to focus only on large pore channels for our further studies of guest molecules in the material3. Furthermore, AIMD simulation proved to be an indispensable tool in the above problems. |
||
16:00 to 16:30 | Recorded talk (APS) | (RLH) ML to Accelerate Experimental Discovery | ||
16:00 to 16:15 | Mini Jose (IISc, India) |
(MLH) Optogenetic modulation of real-time nanoscale dynamics of HCN channels using photoactivated adenylyl cyclases Adenosine 30,50-cyclic monophosphate (cAMP) is a key second messenger that activates several signal transduction pathways in eukaryotic cells. Alteration of basal levels of cAMP is known to activate protein kinases, regulate phosphodiesterases and modulate the activity of ion channels such as Hyper polarization-activated cyclic nucleotide gated channels (HCN). Recent advances in optogenetics have resulted in the availability of novel genetically encoded molecules with the capability to alter cytoplasmic profiles of cAMP with unprecedented spatial and temporal precision. Using single molecule based super-resolution microscopy and different optogenetic modulators of cellular cAMP in both live and fixed cells, we illustrate a novel paradigm to report alteration in nanoscale confinement of ectopically expressed HCN channels. We characterized the efficacy of cAMP generation using ensemble photoactivation of different optogenetic modulators. Then we demonstrate that local modulation of cAMP alters the exchange of membrane bound HCN channels with its nanoenvironment. Additionally, using high density single particle tracking in combination with both acute and chronic optogenetic elevation of cAMP in the cytoplasm, we show that HCN channels are confined to sub 100 nm sized functional domains on the plasma membrane. The nanoscale properties of these domains along with the exchange kinetics of HCN channels in and out of these molecular zones are altered upon temporal changes in the cytoplasmic cAMP. Using HCN2 point mutants and a truncated construct of HCN2 with altered sensitivity to cAMP, we confirmed these alterations in lateral organization of HCN2 to be specific to cAMP binding. Thus, combining these advanced non-invasive paradigms, we report a cAMP dependent ensemble and single particle behavior of HCN channels mediated by its cyclic nucleotide binding domain, opening innovative ways to dissect biochemical pathways at the nanoscale and real-time in living cells. Reference |
||
16:15 to 16:30 | Premchand Rajeev (IISc, India) | (MLH) Nanoscale segregation of synaptic scaffolding protein SAP97 follows first order phase transitions | ||
16:30 to 16:45 | Ayan Roychowdhury (NCBS, India) |
(MLH) Active Force Patterning in a Mixture of Contractile Stresslets The diversity of molecular mechanotransducers, e.g., various forms of myosins, that act as force generators and sensors in cells and tissues, exhibit different levels of contractility and turnover and different cellular localisations. This results in a patterning of nonequilibrium forces within the cell which in turn drives cellular organisation. As a precursor of such active force patterning, here we study the spontaneous segregation of a mixture of contractile stresslets on an elastomer with strain dependent turnover. In contrast to conventional segregation driven by gradients in chemical potential, the spontaneous segregation observed here is driven by elastic stress dissipation. Further, unlike the usual dynamics of coarsening, the linearly segregated state evolves into well separated, singular configurations of contractile stresslets at late times. In addition, the dynamics reveals striking nonreciprocal features, such as travelling waves and a swap phase that breaks time translation. This active stress driven segregation coupled with preferential wetting to substrates has implications for the differential cellular localisations of the diverse myosin species. |
||
16:30 to 16:45 | Ashutosh Kumar Verma (IISc, India) |
(RLH) Combined First-Principles and Classical Modeling of hBN-Water Interfaces: How Surface Roughness and Defects Modulate Wettability and Friction Hexagonal boron nitride (hBN), the inorganic analogue of graphene, is currently being explored for several applications, including membrane separation devices and sensors. In these applications, wettability and friction are crucial interfacial phenomena and understanding them is vital for designing devices for seawater desalination and osmotic power harvesting. In this regard, the water contact angle (WCA) and water slip length (WSL) are the fundamental properties measured in experimental investigations of hBN–water interfaces. Although, in reality, 2D materials contain defects such as vacancies and exposed edges, so far, studies have not considered the effect of defects on the WCA and WSL on hBN surfaces. In this talk, we discuss the wetting and frictional behavior of monolayer and bulk hBN with atomic-scale defects and roughness. To this end, first-principles density functional theory (DFT) and classical-mechanical molecular dynamics simulations are used to calculate the charge distribution inside the hBN nanosheets and to characterize wetting and frictional properties of hBN, respectively. We consider six different topologies of defects in hBN – the B, N, BN, B2N, and B3N vacancy defects, as well as exposed zigzag edges – and also study the effect of the defect concentration on the WCA and WSL to investigate more realistically the interfacial properties of defective hBN. We find that defects at a lower concentration of 0.082 nm-2 no longer affect the wetting properties of hBN surfaces, although they do affect the frictional properties of the surface. On the contrary, exposed edges have a significant effect on the WCA and WSL even at this lower concentration, leading to a notable drop in both quantities, and leading to excellent agreement with experimental data. In summary, monolayer roughness, but not defects, can explain the experimentally observed water contact angle of 66° measured on freshly cleaved, uncontaminated hBN. Not only that, monolayer roughness can explain the measured water slip length of ~1 nm on hBN. Overall, our results indicate the importance of considering realistic models of hBN nanosheets with defects and surface roughness in simulations of water purifcation and energy harvesting applications. |
||
16:45 to 17:00 | Mohammad Arsalan Ashraf (RRI, India) |
(MLH) Contractile force generation in membrane nano-tubes pulled from axonal shaft Neurons while in the growth phase generate membrane nano-tubes, called filopodia, from their growth cone. Filopodia are highly dynamic structures generated by a polar assembly of bundled actin filaments, growth and retraction of filopodia in nerve growth-cone provides motility to neurons. They can sense chemical or mechanical cues and preferential formation of multiple filopodia in a given direction, guided by chemical or other gradients, can guide nerve growth-cone. Filopodia are shown to generate contractile force but the mechanism by which this pulling force is generated is not known. Measuring force in membrane nano-tubes pulled from axonal shaft can help us understand the fiopodia dynamics. These artificially pulled membrane tethers show active load and fail behavior, when they are held for longer times, and it can serve as a model system to study the mechanism for filopodia force generation. We discuss the experiments of pulling membrane tethers from axonal shafts using optical tweezers, our experiments along with a suggested toy model give some insight into the microscopic details about how membrane nano-tubes might generate contractile force. |
||
16:45 to 17:00 | Raju Kumar Biswas (JNCASR, India) |
(RLH) Modeling Pseudo-ternary Phase of Ag2Se For Realization of n-type Thermoelectrics with High zT Recent developments in the thermoelectric application of Ag2Se, which simultaneously shows high mobility and low thermal conductivity, are motivated by innovative thermoelectric theories. However, the relatively narrow bandgap prevents it from achieving a high thermoelectric figure of merit, zT, at room temperature. In this regard, the Rashba effect, spin-dependent band splitting, shows a new direction toward the enhancement of thermoelectric performance. Herein, we investigate the Rashba effect in Ag2Se, which originated mainly due to Te-doping, which provides a unique mechanism for tuning thermoelectric power factors. Additionally, the amorphous limit of lattice thermal conductivity can be achieved via the engineering of configurational entropy dominated by point defect scattering in the pseudo-ternary phase. Density functional theory calculations also reveal that sulfur atoms are locally off-centered, resulting in localized soft optical phonons coupled with acoustics phonons that reduce lattice thermal conductivity. Finally, a benchmark zT~2.1 at 400 K, originated from a combination of Rashba effect along with entropic effect, can provide a new area of research in the thermoelectrics domain. Since, Ag2Se, Ag2Se0.5Te0.5 have already been synthesized, we believe the phase Ag2Se0.5Te0.25S0.25 system will soon be realized experimentally. References: |
||
17:00 to 17:15 | Kishore Hari (IISc, India) |
(MLH) Patterns in complexity: teams of nodes in regulatory networks lead to a robust phenotypic landscape in Epithelial-Mesenchymal plasticity Elucidating the principles of cellular decision-making is of fundamental importance. These decisions are often orchestrated by underlying regulatory networks. While we understand the dynamics of simple network motifs, how do large networks lead to a limited number of phenotypes, despite their complexity, remains largely elusive. Here, we investigate five different networks governing epithelial-mesenchymal plasticity and identified a latent design principle in their topology that limits their phenotypic repertoire – the presence of two “teams” of nodes engaging in a mutually inhibitory feedback loop, forming a toggle switch. These teams are specific to these networks and directly shape the phenotypic landscape and consequently the frequency and stability of terminal phenotypes vs. the intermediary ones. Our analysis reveals that network topology alone can contain information about phenotypic distributions it can lead to, thus obviating the need to simulate them. We unravel topological signatures that can drive the canalization of cell fates during diverse decision-making processes |
||
17:00 to 17:15 | Nikhilesh Maity (IISc, India) |
(RLH) Stacking-Order-Driven Optical Properties and Layer Parity Dependent Phonon Modes in ReS2 Two-dimensional (2D) transition metal dichalcogenides (TMDs) have drawn immense attention from the scientific community because of their spin, valley, optoelectronics, and catalytic applications, making them unique from their bulk counterparts. Recently, among all the TMDs, rhenium disulfide (ReS 2 ) has begun to attract much attention due to its extremely weak interlayer coupling strength and anisotropic properties. The weaker interlayer coupling strength leads to persisting the monolayer behavior of ReS 2 till its bulk phase and resilient its electronic and optical properties with applied external pressure. However, on the other hand, weaker interlayer coupling strength makes it challenging to determine the stacking order of multilayer or bulk structure of ReS 2 . Using first principles density functional theory (DFT) calculations, we have identified two distinct stacking orders (AA & AB) of bilayer ReS 2 , which correspond to two local energy minima of the potential energy surface, resolving the prevailing discrepancies in the literature. Further, we have revealed the stacking-order-driven differences in vibrational and optical properties of bilayer ReS 2 using density functional perturbation theory (DFPT) and GW-Bethe-Salpeter equation (BSE) simulations, respectively. In addition to weak interlayer coupling strength, another uniqueness of ReS 2 lies in its distorted 1T triclinic crystal structure where the additional d valence electrons of Re atoms form zigzag Re chains parallel to the b-axis, drastically reducing its symmetry and making its electronic and optical properties anisotropic. The low triclinic crystal symmetry and in-plane anisotropy in the structure make it challenging to resolve its absolute characteristics, like the origin of extra phonon modes Raman spectra beyond three layer of ReS 2 . We have demonstrated that the observed additional phonon modes are driven by unique layer-parity-dependent splitting of Eg 1 mode and inversion symmetry breaking. The findings underscore the stacking-order driven electronic, vibrational and optical properties of ReS 2 , mediate many seemingly contradictory results in the literature, and open up an opportunity to engineer electronic devices with new functionalities by manipulating the stacking order. |
||
17:15 to 17:30 | Anton Iyer (NCBS, India) |
(MLH) Predicting cancer cell population dynamics using single cell data Isogenic cancer cells develop heterogeneity via a variety of factors like epigenetics, gene expression fluctuations, etc. These differences amongst genetically identical cells lead to divergent cell fate outcomes when subjected to drugs such as chemotherapies. It is well established that only a fraction of these cells die depending on the concentration of the drug administered, a phenomenon known as ‘fractional killing’. However, how this fractional killing effect translates to population growth dynamics of cancer cells is poorly understood. Exponential growth models of cancer cell populations relate the net growth rate of a population to the single cell division and death times. Such models have been used widely in quantifying cell proliferation in response to drugs and have also been used to design improved clinical strategies for cancer treatment. With the improvement in single cell time-lapse imaging techniques, it has become possible to try and predict the population dynamics from single cell data. For U2OS cells proliferating in the absence of any drug, the deviations between theoretical predictions of growth rate and regression fits to the data could be minimized significantly provided the model incorporated a factor of synchronicity in cell age at time of seeding. Surprisingly however, the same theories failed to predict the observed growth rates after administration of the chemotherapeutic agent cisplatin. The underlying distributions of cell death and division times, inferred using a likelihood-based approach to account for competing fates, were found to vary insignificantly on increasing the concentration of the drug even though the net population growth dramatically changed. These surprising results could not be explained using the traditional theories connecting single cell data to population dynamics, revealing a lacuna in the current understanding of how drug resistant cancer cells contribute to cancer cell proliferation. My future work will focus on identifying other essential single cell parameters that are required for developing a quantitative understanding of the population dynamics of drug-treated cancer cells. |
||
17:15 to 17:45 | Nikhil Malviya (IISc, India) |
(RLH) Phonon transport in ultrahigh thermal conductivity materials beyond the relaxation time approximation In electrical insulators, heat is carried by the quantized collective lattice vibrations called phonons. Resistance to heat flow in these materials is caused by phonon scattering processes. Thermal phonon transport in these materials is governed by the semi-classical Boltzmann Transport Equation (BTE). Solutions of the BTE are commonly derived assuming the validity of relaxation time approximation (RTA), where all phonon scattering events are assumed to be momentum-dissipative in nature. While the RTA-based BTE solution describes the heat flow in several materials reasonably well, it fails to capture the ultrahigh thermal conductivity and the exceptional phonon transport properties of materials like diamond and boron nitride. Here we present the solutions of the BTE without the RTA for phonon transport through these ultrahigh thermal conductivity materials, and demonstrate that accurately distinguishing momentum-conserving (Normal) and momentum-dissipative (Umklapp) scattering events in our formulation is crucial to correctly predict their thermal transport properties. This work is supported by the DST-SERB Core Research Grant (CRG/2020/006166) and Prime Minister Research Fellowship (192002-2069). |
||
17:30 to 17:45 | Rajeswari Appadurai (IISc, India) |
(MLH) Simplified interpretation of complex biomolecular ensemble data using dimensionality reduction techniques: From unifying quality assessment to data-tailored application In an era of data revolution, techniques to obtain simplified physical intuition from large complex datasets is equally essential as that of the generation of data in itself. Dimensionality reduction (DR) techniques are often employed to extract values from complex datasets and have found wide applications across several fields, in which our primary interest is in its application in protein conformational ensemble analyses. However, the kind of DR technique (among several linear and non-linear methods) appropriate for sensible data applications varies depending on the context of the data. In this work we implement a unifying quality assessment framework for evaluating various DR projections based on the mismatches and inaccuracies in both distance rankings and distance relations of the embedded points with respect to that in the original high dimensional space. We apply the framework on very long trajectories of fast folding proteins and intrinsically disordered proteins (IDP). We find that, the automated metrics data from our quality assessment framework correlates well with interpretations obtained using domain knowledge. While there is no general approach that fits for all trajectories, we see that the t-distributed stochastic neighbour embedding (TSNE) consistently performs well in terms of the rank-based metrics though preserving the distance relations is subjective to the hyper-parameter tuning. Indeed the method is designed to preserve the local structure than the global relations and thus can provide a very informative visualization of heterogeneity in the data. We show that how this particular aspect of TSNE can be exploited for clustering an IDP ensemble into distinct conformational sub-populations that further enlighten about their unique topological design and ligand binding abilities which are otherwise hard to obtain. |
||
17:45 to 18:00 | Rohit Kumar Rohj (IISc, India) | (RLH) Reduction in band gap with retention of ferroelectricity through heterovalent codoping in BaTiO3 | ||
17:45 to 18:00 | -- | (MLH) Discussions |