Dataset Information

Branching principles of animal and plant networks identified by combining extensive data, machine learning and modelling.

ABSTRACT: Branching in vascular networks and in overall organismic form is one of the most common and ancient features of multicellular plants, fungi and animals. By combining machine-learning techniques with new theory that relates vascular form to metabolic function, we enable novel classification of diverse branching networks-mouse lung, human head and torso, angiosperm and gymnosperm plants. We find that ratios of limb radii-which dictate essential biologic functions related to resource transport and supply-are best at distinguishing branching networks. We also show how variation in vascular and branching geometry persists despite observing a convergent relationship across organisms for how metabolic rate depends on body mass.

SUBMITTER: Brummer AB

PROVIDER: S-EPMC7879751 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Branching principles of animal and plant networks identified by combining extensive data, machine learning and modelling.

Brummer Alexander B AB Lymperopoulos Panagiotis P Shen Jocelyn J Tekin Elif E Bentley Lisa P LP Buzzard Vanessa V Gray Andrew A Oliveras Imma I Enquist Brian J BJ Savage Van M VM

Journal of the Royal Society, Interface 20210106 174

Branching in vascular networks and in overall organismic form is one of the most common and ancient features of multicellular plants, fungi and animals. By combining machine-learning techniques with new theory that relates vascular form to metabolic function, we enable novel classification of diverse branching networks-mouse lung, human head and torso, angiosperm and gymnosperm plants. We find that ratios of limb radii-which dictate essential biologic functions related to resource transport and ...[more]

PMID: 33402023

Similar Datasets

Project description:Among ?-transglucosylases from Glycoside-Hydrolase family 70, the ?N123-GB-CD2 enzyme derived from the bifunctional DSR-E from L. citreum NRRL B-1299 is particularly interesting as it was the first described engineered Branching Sucrase, not able to elongate glucan polymers from sucrose substrate. The previously reported overall structural organization of this multi-domain enzyme is an intricate U-shape fold conserved among GH70 enzymes which showed a certain conformational variability of the so-called domain V, assumed to play a role in the control of product structures, in available X-ray structures. Understanding the role of functional dynamics on enzyme reaction and substrate recognition is of utmost interest although it remains a challenge for biophysical methods. By combining long molecular dynamics simulation (1?s) and multiple analyses (NMA, PCA, Morelet Continuous Wavelet Transform and Cross Correlations Dynamics), we investigated here the dynamics of ?N123-GB-CD2 alone and in interaction with sucrose substrate. Overall, our results provide the detailed picture at atomic level of the hierarchy of motions occurring along different timescales and how they are correlated, in agreement with experimental structural data. In particular, detailed analysis of the different structural domains revealed cooperative dynamic behaviors such as twisting, bending and wobbling through anti- and correlated motions, and also two structural hinge regions, of which one was unreported. Several highly flexible loops surrounding the catalytic pocket were also highlighted, suggesting a potential role in the acceptor promiscuity of ?N123-GBD-CD2. Normal modes and essential dynamics underlined an interesting two-fold dynamic of the catalytic domain A, pivoting about an axis splitting the catalytic gorge in two parts. The comparison of the conformational free energy landscapes using principal component analysis of the enzyme in absence or in presence of sucrose, also revealed a more harmonic basin when sucrose is bound with a shift population of the bending mode, consistent with the substrate binding event.

Project description:Introduction: Hybrid modeling combining First-Principles with machine learning is becoming a pivotal methodology for Biopharma 4.0 enactment. Chinese Hamster Ovary (CHO) cells, being the workhorse for industrial glycoproteins production, have been the object of several hybrid modeling studies. Most previous studies pursued a shallow hybrid modeling approach based on three-layered Feedforward Neural Networks (FFNNs) combined with macroscopic material balance equations. Only recently, the hybrid modeling field is incorporating deep learning into its framework with significant gains in descriptive and predictive power. Methods: This study compares, for the first time, deep and shallow hybrid modeling in a CHO process development context. Data of 24 fed-batch cultivations of a CHO-K1 cell line expressing a target glycoprotein, comprising 30 measured state variables over time, were used to compare both methodologies. Hybrid models with varying FFNN depths (3-5 layers) were systematically compared using two training methodologies. The classical training is based on the Levenberg-Marquardt algorithm, indirect sensitivity equations and cross-validation. The deep learning is based on the Adaptive Moment Estimation Method (ADAM), stochastic regularization and semidirect sensitivity equations. Results and conclusion: The results point to a systematic generalization improvement of deep hybrid models over shallow hybrid models. Overall, the training and testing errors decreased by 14.0% and 23.6% respectively when applying the deep methodology. The Central Processing Unit (CPU) time for training the deep hybrid model increased by 31.6% mainly due to the higher FFNN complexity. The final deep hybrid model is shown to predict the dynamics of the 30 state variables within the error bounds in every test experiment. Notably, the deep hybrid model could predict the metabolic shifts in key metabolites (e.g., lactate, ammonium, glutamine and glutamate) in the test experiments. We expect deep hybrid modeling to accelerate the deployment of high-fidelity digital twins in the biopharma sector in the near future.

Dataset Information

Branching principles of animal and plant networks identified by combining extensive data, machine learning and modelling.

Publications

Branching principles of animal and plant networks identified by combining extensive data, machine learning and modelling.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets