-
Top-down machine learning of coarse-grained protein force-fields
Authors:
Carles Navarro,
Maciej Majewski,
Gianni de Fabritiis
Abstract:
Develo** accurate and efficient coarse-grained representations of proteins is crucial for understanding their folding, function, and interactions over extended timescales. Our methodology involves simulating proteins with molecular dynamics and utilizing the resulting trajectories to train a neural network potential through differentiable trajectory reweighting. Remarkably, this method requires…
▽ More
Develo** accurate and efficient coarse-grained representations of proteins is crucial for understanding their folding, function, and interactions over extended timescales. Our methodology involves simulating proteins with molecular dynamics and utilizing the resulting trajectories to train a neural network potential through differentiable trajectory reweighting. Remarkably, this method requires only the native conformation of proteins, eliminating the need for labeled data derived from extensive simulations or memory-intensive end-to-end differentiable simulations. Once trained, the model can be employed to run parallel molecular dynamics simulations and sample folding events for proteins both within and beyond the training distribution, showcasing its extrapolation capabilities. By applying Markov State Models, native-like conformations of the simulated proteins can be predicted from the coarse-grained simulations. Owing to its theoretical transferability and ability to use solely experimental static structures as training data, we anticipate that this approach will prove advantageous for develo** new protein force fields and further advancing the study of protein dynamics, folding, and interactions.
△ Less
Submitted 10 October, 2023; v1 submitted 20 June, 2023;
originally announced June 2023.
-
Modelling the effect of vascular status on tumour evolution and outcome after thermal therapy
Authors:
Jesús J. Bosque,
Gabriel F. Calvo,
María Cruz Navarro
Abstract:
Microscale oxygenation plays a prominent role in tumour progression. Spatiotemporal variability of oxygen distribution in the tumour microenvironment contributes to cellular heterogeneity and to the emergence of normoxic and hypoxic populations. Local levels of oxygen strongly affect the response of tumours to the administration of different therapeutic modalities and, more generally, to the pheno…
▽ More
Microscale oxygenation plays a prominent role in tumour progression. Spatiotemporal variability of oxygen distribution in the tumour microenvironment contributes to cellular heterogeneity and to the emergence of normoxic and hypoxic populations. Local levels of oxygen strongly affect the response of tumours to the administration of different therapeutic modalities and, more generally, to the phenomenon of resistance to treatments. Several interventions have been proposed to improve tumour oxygenation, being the elevation of the local temperature (hyperthermia) an important one. While other factors such as the metabolic activity have to be considered, the proficiency of the tumour vascular system is a key factor both for the tissue oxygenation and for its temperature maps. Consequently, the interplay of these factors has attracted considerable attention from the mathematical modelling perspective. Here we put forward a transport-based system of partial differential equations aimed at describing the dynamics of healthy and tumour cell subpopulations at the microscale in a region placed between two blood vessels. By using this model with diverse flow conditions, we analyse the oxygen and temperature profiles that arise in different scenarios of vascular status, both during free progression and under thermal therapy. We find that low oxygen levels are associated to elevations of temperature in locations preferentially populated by hypoxic cells, and hyperthermia-induced cell death, being strongly dependent on blood flow, would only appear under highly disrupted conditions of the local vasculature. This results in a noticeable effect of heat on hypoxic cells. Additionally, when pronounced cell death occurs, it is followed by a significant increase in the oxygen levels.
△ Less
Submitted 10 January, 2022;
originally announced January 2022.
-
Reliable scaling of Position Weight Matrices for binding strength comparisons between transcription factors
Authors:
Xiaoyan Ma,
Daphne Ezer,
Carmen Navarro,
Boris Adryan
Abstract:
Scoring DNA sequences against Position Weight Matrices (PWMs) is a widely adopted method to identify putative transcription factor binding sites. While common bioinformatics tools produce scores that can reflect the binding strength between a specific transcription factor and the DNA, these scores are not directly comparable between different transcription factors. Here, we provide two different w…
▽ More
Scoring DNA sequences against Position Weight Matrices (PWMs) is a widely adopted method to identify putative transcription factor binding sites. While common bioinformatics tools produce scores that can reflect the binding strength between a specific transcription factor and the DNA, these scores are not directly comparable between different transcription factors. Here, we provide two different ways to find the scaling parameter $λ$ that allows us to infer binding energy from a PWM score. The first approach uses a PWM and background genomic sequence as input to estimate $λ$ for a specific transcription factor, which we applied to show that $λ$ distributions for different transcription factor families correspond with their DNA binding properties. Our second method can reliably convert $λ$ between different PWMs of the same transcription factor, which allows us to directly compare PWMs that were generated by different approaches. These two approaches provide consistent and computationally efficient ways to scale PWMs scores and estimate transcription factor binding sites strength.
△ Less
Submitted 17 March, 2015;
originally announced March 2015.
