Search

The determination of permeability is an example of many geological problems where laboratory-measured data is expensive and limited in quantity. We related permeability values to well logs. We used neural networks trained both with the popular back-propagation algorithm and with a genetic algorithm. The genetic training produced smaller errors and better generalization than back-propagation training on the same network topology. The cost includes greater average computation time as well as greater variation in computation time for the genetic training. The genetic training is robust and not sensitive to selectionof the crossover and mutation parameters. [ABSTRACT FROM AUTHOR]

In petroleum geology, lithofacies information is important for estimating porosity and permeability values from wireline logs at the un-cored well intervals; however, predicting lithofacies from logs is notan easy task. The results of lithofacies prediction from wireline log signals were compared using two different supervised classificationmethods: backpropagation neural network (BPNN) and a simplified version of Fuzzy ARTMAP called SFAM. The SFAM method gives results similar to those of BPNNs, but does not suffer BPNN's problems of excessivetraining time and the need for prior specification of network topology. If training time and the effort required to fine-tune BPNN parameters are acceptable, in some cases BPNN can provide significantly improved performance. However, to achieve this performance will generally require some degree of skill and a process of trial and error. Porosity and permeability predictions can be estimated by BPNNs; however,SFAM is currently unable to perform this task, and this requires further study. [ABSTRACT FROM AUTHOR]

The proteome balance constraint in metabolic flux balance analysis asserts that the proteome is constructed by ribosomes, which themselves contain many proteins. This leads to a fundamental question of optimal allocation of limited proteome among different pools of enzymes, which include ribosomes themselves. However, recent work points to additional constraints imposed by the cell geometry. In this paper we deduce the proteogeometric constraintπ¯ A =π A +θπ L /π P , where π A , π P and π L are the proteomic fractions allocated to the cell surface area, protein synthesis and cell membrane phospholipids synthesis and π¯ A and θ are constants imposed by geometry of the cell. We illustrate the relevance of this constraint using a reduced model of cell metabolism, illuminating the interplay between cell metabolism and cell geometry., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2025 Elsevier Ltd. All rights reserved.)

Most existing graph neural networks (GNNs) learn node embeddings using the framework of message passing and aggregation. Such GNNs are incapable of learning relative positions between graph nodes within a graph. To empower GNNs with the awareness of node positions, some nodes are set as anchors. Then, using the distances from a node to the anchors, GNNs can infer relative positions between nodes. However, position-aware GNNs (P-GNNs) arbitrarily select anchors, leading to compromising position awareness and feature extraction. To eliminate this compromise, we demonstrate that selecting evenly distributed and asymmetric anchors is essential. On the other hand, we show that choosing anchors that can aggregate embeddings of all the nodes within a graph is NP-complete. Therefore, devising efficient optimal algorithms in a deterministic approach is practically not feasible. To ensure position awareness and bypass NP-completeness, we propose position-sensing GNNs (PSGNNs), learning how to choose anchors in a backpropagatable fashion. Experiments verify the effectiveness of PSGNNs against state-of-the-art GNNs, substantially improving performance on various synthetic and real-world graph datasets while enjoying stable scalability. Specifically, PSGNNs on average boost area under the curve (AUC) more than 14% for pairwise node classification and 18% for link prediction over the existing state-of-the-art position-aware methods. Our source code is publicly available at: https://github.com/ZhenyueQin/PSGNN.

In response to the global need for efficient early diagnosis of Autism Spectrum Disorder (ASD), this paper bridges the gap between traditional, time-consuming diagnostic methods and potential automated solutions. We propose a multi-atlas deep ensemble network, MADE-for-ASD, that integrates multiple atlases of the brain's functional magnetic resonance imaging (fMRI) data through a weighted deep ensemble network. Our approach integrates demographic information into the prediction workflow, which enhances ASD diagnosis performance and offers a more holistic perspective on patient profiling. We experiment with the well-known publicly available ABIDE (Autism Brain Imaging Data Exchange) I dataset, consisting of resting state fMRI data from 17 different laboratories around the globe. Our proposed system achieves 75.20% accuracy on the entire dataset and 96.40% on a specific subset - both surpassing reported ASD diagnosis accuracy in ABIDE I fMRI studies. Specifically, our model improves by 4.4 percentage points over prior works on the same amount of data. The model exhibits a sensitivity of 82.90% and a specificity of 69.70% on the entire dataset, and 91.00% and 99.50%, respectively, on the specific subset. We leverage the F-score to pinpoint the top 10 ROI in ASD diagnosis, such as precuneus and anterior cingulate/ventromedial. The proposed system can potentially pave the way for more cost-effective, efficient and scalable strategies in ASD diagnosis. Codes and evaluations are publicly available at https://github.com/hasan-rakibul/MADE-for-ASD., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

Elucidating the emergent dynamics of cellular differentiation networks is crucial to understanding cell-fate decisions. Toggle switch - a network of mutually repressive lineage-specific transcription factors A and B - enables two phenotypes from a common progenitor: (high A, low B) and (low A, high B). However, the dynamics of networks enabling differentiation of more than two phenotypes from a progenitor cell has not been well-studied. Here, we investigate the dynamics of a four-node network A, B, C, and D inhibiting each other, forming a toggle tetrahedron. Our simulations show that this network is multistable and predominantly allows for the co-existence of six hybrid phenotypes where two of the nodes are expressed relatively high as compared to the remaining two, for instance (high A, high B, low C, low D). Finally, we apply our results to understand naïve CD4 + T cell differentiation into Th1, Th2, Th17 and Treg subsets, suggesting Th1/Th2/Th17/Treg decision-making to be a two-step process., (© 2024. The Author(s).)

In this review paper we summarize a recent progress on the problem of describing range of dynamics supported by a network. We show that there is natural connection between network models consisting of collections of multivalued monotone boolean functions and ordinary differential equations models. We show how to construct such collections and use them to answer questions about prevalence of cellular phenotypes that correspond to equilibria of network models., (© 2024. The Author(s).)

A metabolic theory is presented for predicting maximum growth rate, overflow metabolism, respiration efficiency, and maintenance energy flux based on the intersection of cell geometry, membrane protein crowding, and metabolism. The importance of cytosolic macromolecular crowding on phenotype has been established in the literature but the importance of surface area has been largely overlooked due to incomplete knowledge of membrane properties. We demonstrate that the capacity of the membrane to host proteins increases with growth rate offsetting decreases in surface area-to-volume ratios (SA:V). This increase in membrane protein is hypothesized to be essential to competitive Escherichia coli phenotypes. The presented membrane-centric theory uses biophysical properties and metabolic systems analysis to successfully predict the phenotypes of E. coli K-12 strains, MG1655 and NCM3722, which are genetically similar but have SA:V ratios that differ up to 30%, maximum growth rates on glucose media that differ by 40%, and overflow phenotypes that start at growth rates that differ by 80%. These analyses did not consider cytosolic macromolecular crowding, highlighting the distinct properties of the presented theory. Cell geometry and membrane protein crowding are significant biophysical constraints on phenotype and provide a theoretical framework for improved understanding and control of cell biology.

We consider two types of models of regulatory network dynamics: Boolean maps and systems of switching ordinary differential equations. Our goal is to construct all models in each category that are compatible with the directed signed graph that describe the network interactions. This leads to consideration of lattice of monotone Boolean functions (MBF), poset of non-degenerate MBFs, and a lattice of chains in these sets. We describe explicit inductive construction of these posets where the induction is on the number of inputs in MBF. Our results allow enumeration of potential dynamic behavior of the network for both model types, subject to practical limitation imposed by the size of the lattice of MBFs described by the Dedekind number., Competing Interests: Declaration of competing interest I declare that I have no known conflict of interest., (Copyright © 2024 Elsevier Inc. All rights reserved.)

Skeleton sequences are lightweight and compact and thus are ideal candidates for action recognition on edge devices. Recent skeleton-based action recognition methods extract features from 3-D joint coordinates as spatial-temporal cues, using these representations in a graph neural network for feature fusion to boost recognition performance. The use of first- and second-order features, that is, joint and bone representations, has led to high accuracy. Nonetheless, many models are still confused by actions that have similar motion trajectories. To address these issues, we propose fusing higher-order features in the form of angular encoding (AGE) into modern architectures to robustly capture the relationships between joints and body parts. This simple fusion with popular spatial-temporal graph neural networks achieves new state-of-the-art accuracy in two large benchmarks, including NTU60 and NTU120, while employing fewer parameters and reduced run time. Our source code is publicly available at: https://github.com/ZhenyueQin/Angular-Skeleton-Encoding.

This article aims to use graphic engines to simulate a large number of training data that have free annotations and possibly strongly resemble to real-world data. Between synthetic and real, a two-level domain gap exists, involving content level and appearance level. While the latter is concerned with appearance style, the former problem arises from a different mechanism, i.e., content mismatch in attributes such as camera viewpoint, object placement and lighting conditions. In contrast to the widely-studied appearance-level gap, the content-level discrepancy has not been broadly studied. To address the content-level misalignment, we propose an attribute descent approach that automatically optimizes engine attributes to enable synthetic data to approximate real-world data. We verify our method on object-centric tasks, wherein an object takes up a major portion of an image. In these tasks, the search space is relatively small, and the optimization of each attribute yields sufficiently obvious supervision signals. We collect a new synthetic asset VehicleX, and reformat and reuse existing the synthetic assets ObjectX and PersonX. Extensive experiments on image classification and object re-identification confirm that adapted synthetic data can be effectively used in three scenarios: training with synthetic data only, training data augmentation and numerically understanding dataset content.

Early development of Drosophila melanogaster (fruit fly) facilitated by the gap gene network has been shown to be incredibly robust, and the same patterns emerge even when the process is seriously disrupted. We investigate this robustness using a previously developed computational framework called DSGRN (Dynamic Signatures Generated by Regulatory Networks). Our mathematical innovations include the conceptual extension of this established modeling technique to enable modeling of spatially monotone environmental effects, as well as the development of a collection of graph theoretic robustness scores for network models. This allows us to rank order the robustness of network models of cellular systems where each cell contains the same genetic network topology but operates under a parameter regime that changes continuously from cell to cell. We demonstrate the power of this method by comparing the robustness of two previously introduced network models of gap gene expression along the anterior-posterior axis of the fruit fly embryo, both to each other and to a random sample of networks with same number of nodes and edges. We observe that there is a substantial difference in robustness scores between the two models. Our biological insight is that random network topologies are in general capable of reproducing complex patterns of expression, but that using measures of robustness to rank order networks permits a large reduction in hypothesis space for highly conserved systems such as developmental networks., Competing Interests: Declaration of competing interest None, (Copyright © 2024 Elsevier Ltd. All rights reserved.)

We introduce two time-delay models of metabolic oscillations in yeast cells. Our model tests a hypothesis that the oscillations occur as multiple pathways share a limited resource which we equate to the number of available ribosomes. We initially explore a single-protein model with a constraint equation governing the total resource available to the cell. The model is then extended to include three proteins that share a resource pool. Three approaches are considered at constant delay to numerically detect oscillations. First, we use a spectral element method to approximate the system as a discrete map and evaluate the stability of the linearized system about its equilibria by examining its eigenvalues. For the second method, we plot amplitudes of the simulation trajectories in 2D projections of the parameter space. We use a history function that is consistent with published experimental results to obtain metabolic oscillations. Finally, the spectral element method is used to convert the system to a boundary value problem whose solutions correspond to approximate periodic solutions of the system. Our results show that certain combinations of total resource available and the time delay, lead to oscillations. We observe that an oscillation region in the parameter space is between regions admitting steady states that correspond to zero and constant production. Similar behavior is found with the three-protein model where all proteins require the same production time. However, a shift in the protein production rates peaks occurs for low available resource suggesting that our model captures the shared resource pool dynamics., (© 2023. The Author(s), under exclusive licence to Society for Mathematical Biology.)