Cell cycle time series gene expression data encoded as cyclic attractors in Hopfield systems

Characterizing Cellular Differentiation Potency and Waddington Landscape via Energy Indicator

The precise characterization of cellular differentiation potency remains an open question, which is fundamentally important for deciphering the dynamics mechanism related to cell fate transition. We quantitatively evaluated the differentiation potency of different stem cells based on the Hopfield neural network (HNN). The results emphasized that cellular differentiation potency can be approximated by Hopfield energy values. We then profiled the Waddington energy landscape of embryogenesis and cell reprogramming processes. The energy landscape at single-cell resolution further confirmed that cell fate decision is progressively specified in a continuous process. Moreover, the transition of cells from one steady state to another in embryogenesis and cell reprogramming processes was dynamically simulated on the energy ladder. These two processes can be metaphorized as the motion of descending and ascending ladders, respectively. We further deciphered the dynamics of the gene regulatory network (GRN) for driving cell fate transition. Our study proposes a new energy indicator to quantitatively characterize cellular differentiation potency without prior knowledge, facilitating the further exploration of the potential mechanism of cellular plasticity.

Network hierarchy and pattern recovery in directed sparse Hopfield networks

Many real-world networks are directed, sparse, and hierarchical, with a mixture of feedforward and feedback connections with respect to the hierarchy. Moreover, a small number of master nodes are often able to drive the whole system. We study the dynamics of pattern presentation and recovery on sparse, directed, Hopfield-like neural networks using trophic analysis to characterize their hierarchical structure. This is a recent method which quantifies the local position of each node in a hierarchy (trophic level) as well as the global directionality of the network (trophic coherence). We show that even in a recurrent network, the state of the system can be controlled by a small subset of neurons which can be identified by their low trophic levels. We also find that performance at the pattern recovery task can be significantly improved by tuning the trophic coherence and other topological properties of the network. This may explain the relatively sparse and coherent structures observed in the animal brain and provide insights for improving the architectures of artificial neural networks. Moreover, we expect that the principles we demonstrate here, through numerical analysis, will be relevant for a broad class of system whose underlying network structure is directed and sparse, such as biological, social, or financial networks.

Prediction of Time Series Gene Expression and Structural Analysis of Gene Regulatory Networks Using Recurrent Neural Networks

Methods for time series prediction and classification of gene regulatory networks (GRNs) from gene expression data have been treated separately so far. The recent emergence of attention-based recurrent neural network (RNN) models boosted the interpretability of RNN parameters, making them appealing for the understanding of gene interactions. In this work, we generated synthetic time series gene expression data from a range of archetypal GRNs and we relied on a dual attention RNN to predict the gene temporal dynamics. We show that the prediction is extremely accurate for GRNs with different architectures. Next, we focused on the attention mechanism of the RNN and, using tools from graph theory, we found that its graph properties allow one to hierarchically distinguish different architectures of the GRN. We show that the GRN responded differently to the addition of noise in the prediction by the RNN and we related the noise response to the analysis of the attention mechanism. In conclusion, this work provides a way to understand and exploit the attention mechanism of RNNs and it paves the way to RNN-based methods for time series prediction and inference of GRNs from gene expression data.

TTK inhibitor promotes radiosensitivity of liver cancer cells through p21

The monopolar spindle 1 ((hMps1/TTK) is a serine/threonine kinase that plays an important role in spindle assembly checkpoint signaling. To explore the possible relationship between TTK inhibition and radiosensitivity, we examined whether TTK inhibition influences cellular susceptibility of radiation. And we further revealed its mechanisms. We found that the expression of TTK was obviously higher in liver cancer tissues compared to the normal liver tissues. Kaplan-Meier Plotter demonstrated that patients with low TTK expression levels had a longer overall survival than patients with high TTK expression levels. TTK inhibitor AZ3146 could simulated liver cancer cells to accumulate in the G2/M phase, which ultimately enhances DNA damage with more γ-H2AX foci and more apoptosis and necrosis induced by radiation, which prompted that TTK inhibition sensitized liver cancer cells to radiation. In addition, TTK inhibition altered cell-cycle progression and exacerbated centrosome abnormalities, resulting in enhanced mitotic catastrophe (MC) induced by radiation in a p21-mediated manner. In this study, we present evidences that the TTK inhibitor promotes the radiosensitivity of liver cancer cells through regulating cell cycle in p21-mediated manner in vitro, indicating that TTK inhibitor may be an attractive radiosensitizer for the patients with liver cancer.

Digital Cell Sorter (DCS): a cell type identification, anomaly detection, and Hopfield landscapes toolkit for single-cell transcriptomics

MOTIVATION

Analysis of singe cell RNA sequencing (scRNA-seq) typically consists of different steps including quality control, batch correction, clustering, cell identification and characterization, and visualization. The amount of scRNA-seq data is growing extremely fast, and novel algorithmic approaches improving these steps are key to extract more biological information. Here, we introduce: (i) two methods for automatic cell type identification (i.e., without expert curator) based on a voting algorithm and a Hopfield classifier, (ii) a method for cell anomaly quantification based on isolation forest, and (iii) a tool for the visualization of cell phenotypic landscapes based on Hopfield energy-like functions. These new approaches are integrated in a software platform that includes many other state-of-the-art methodologies and provides a self-contained toolkit for scRNA-seq analysis.

RESULTS

We present a suite of software elements for the analysis of scRNA-seq data. This Python-based open source software, Digital Cell Sorter (DCS), consists in an extensive toolkit of methods for scRNA-seq analysis. We illustrate the capability of the software using data from large datasets of peripheral blood mononuclear cells (PBMC), as well as plasma cells of bone marrow samples from healthy donors and multiple myeloma patients. We test the novel algorithms by evaluating their ability to deconvolve cell mixtures and detect small numbers of anomalous cells in PBMC data.

AVAILABILITY

The DCS toolkit is available for download and installation through the Python Package Index (PyPI). The software can be deployed using the Python import function following installation. Source code is also available for download on Zenodo: DOI 10.5281/zenodo.2533377.

SUPPLEMENTARY INFORMATION

Supplemental Materials are available at PeerJ online.

Modeling disease progression in Multiple Myeloma with Hopfield networks and single-cell RNA-seq

Associative memories in Hopfield's neural networks are mapped to gene expression pattern to model different paths of disease progression towards Multiple Myeloma (MM). The model is built using single cell RNA-seq data from bone marrow aspirates of MM patients as well as patients diagnosed with Monoclonal Gammopathy of Undetermined Significance (MGUS) and Smoldering Multiple Myeloma (SMM), two medical conditions that often progress to full MM.

RESULTS

We identify different clusters of MGUS, SMM, and MM cells, map them to Hopfield associative memory patterns, and model the dynamics of transition between the different patterns. The model is then used to identify genes that are differentialy expressed across different MM stages and whose simultaneous inhibition is associated to a delayed disease progression.

Hope4Genes: a Hopfield-like class prediction algorithm for transcriptomic data

After its introduction in 1982, the Hopfield model has been extensively applied for classification and pattern recognition. Recently, its great potential in gene expression patterns retrieval has also been shown. Following this line, we develop Hope4Genes a single-sample class prediction algorithm based on a Hopfield-like model. Differently from previous works, we here tested the performances of the algorithm for class prediction, a task of fundamental importance for precision medicine and therapeutic decision-making. Hope4Genes proved better performances than the state-of-art methodologies in the field independently of the size of the input dataset, its profiling platform, the number of classes and the typical class-imbalance present in biological data. Our results provide encoraging evidence that the Hopfield model, together with the use of its energy for the estimation of the false discoveries, is a particularly promising tool for precision medicine.

Genetic Markers Can Predict Chondrogenic Differentiation Potential in Bone Marrow-Derived Mesenchymal Stromal Cells

The precise predictions of the differentiation direction and potential of mesenchymal stromal cells (MSCs) are an important key to the success of regenerative medicine. The expression levels of fate-determining genes may provide tools for predicting differentiation potential. The expression levels of 95 candidate marker genes and glycosaminoglycan (GAG) contents after chondrogenic induction in 10 undifferentiated ilium and 5 jaw MSC cultures were determined, and their correlations were analyzed. The expression levels of eight genes before the induction of chondrogenic MSC differentiation were significantly correlated with the GAG levels after induction. Based on correlation patterns, the eight genes were classified into two groups: group 1 genes (AURKB, E2F1, CDKN2D, LIF, and ACLY), related to cell cycle regulation, and group 2 genes (CD74, EFEMP1, and TGM2), involved in chondrogenesis. The expression levels of the group 2 genes were significantly correlated with the ages of the cell donors. The expression levels of CDKN2D, CD74, and TGM2 were >10-fold higher in highly potent MSCs (ilium MSCs) than in MSCs with limited potential (jaw MSCs). Three-dimensional (3D) scatter plot analyses of the expression levels of these genes showed reduced variability between donors and confirmed predictive potential. These data suggest that group 2 genes are involved in age-dependent decreases in the chondrogenic differentiation potential of MSCs, and combined 3D analyses of the expression profiles of three genes, including two group 2 genes, were predictive of MSC differentiation potential.