Inferring latent temporal progression and regulatory networks from cross-sectional transcriptomic data of cancer samples

Baseline atherogenic index of plasma and its trajectory predict onset of type 2 diabetes in a health screened adult population: a large longitudinal study

BACKGROUND

The Atherogenic Index of Plasma (AIP) is a novel biomarker for assessing the severity of atherosclerosis and has been shown to be closely associated with the risk of Type 2 Diabetes Mellitus (T2DM). However, no prospective cohort study has comprehensively evaluated both the immediate risk stratification through baseline AIP and the long-term risk assessment through multi-time point AIP trajectories in health screened adults in relation to T2DM risk.

METHODS

This longitudinal study included data from 42,850 participants who underwent health check-ups at Henan Provincial People's Hospital between January 2018 and August 2024. AIP was calculated as the logarithm of the ratio of triglycerides (TG) to high-density lipoprotein cholesterol (HDL-C). The Kaplan-Meier method was employed to analyze the incidence of T2DM across different AIP groups. A Cox model with restricted cubic splines assessed the dose-response relationship between AIP and T2DM risk, while latent class trajectory models (LCTM) evaluated the trends of AIP over multiple time points. Cox proportional hazards models were used to examine the relationship between baseline AIP quartiles, AIP trajectories, and T2DM risk.

RESULTS

During an average follow-up of 47.95 months, 3,228 participants developed T2DM. Stratifying by baseline AIP quartiles revealed that higher AIP levels were associated with an increased risk of T2DM. Compared to the lowest quartile, the highest quartile had a hazard ratio (HR) of 2.10 (95% CI: 1.74, 2.53). The LCTM identified three trajectory patterns for AIP: with the low-stable group as the reference, the medium-stable and high-stable groups had HRs of 1.72 (95% CI: 1.50, 1.96) and 2.50 (95% CI: 2.06, 3.03), respectively, indicating a significantly elevated risk of T2DM (P < 0.05).

CONCLUSION

Elevated baseline AIP levels, medium stable trajectories and high stable trajectories are associated with an increased risk of T2DM in health screened adults.

Constrained Pseudo-Time Ordering for Clinical Transcriptomics Data

Time series RNASeq studies can enable understanding of the dynamics of disease progression and treatment response in patients. They also provide information on biomarkers, activated and repressed pathways, and more. While useful, data from multiple patients is challenging to integrate due to the heterogeneity in treatment response among patients, and the small number of timepoints that are usually profiled. Due to the heterogeneity among patients, relying on the sampled time points to integrate data across individuals is challenging and does not lead to correct reconstruction of the response patterns. To address these challenges, we developed a new constrained based pseudo-time ordering method for analyzing transcriptomics data in clinical and response studies. Our method allows the assignment of samples to their correct placement on the response curve while respecting the individual patient order. We use polynomials to represent gene expression over the duration of the study and an EM algorithm to determine parameters and locations. Application to four treatment response datasets shows that our method improves on prior methods and leads to accurate orderings that provide new biological insight on the disease and response.

CSER: a gene regulatory network construction method based on causal strength and ensemble regression

Introduction

Gene regulatory networks (GRNs) reveal the intricate interactions between and among genes, and understanding these interactions is essential for revealing the molecular mechanisms of cancer. However, existing algorithms for constructing GRNs may confuse regulatory relationships and complicate the determination of network directionality.

Methods

We propose a new method to construct GRNs based on causal strength and ensemble regression (CSER) to overcome these issues. CSER uses conditional mutual inclusive information to quantify the causal associations between genes, eliminating indirect regulation and marginal genes. It considers linear and nonlinear features and uses ensemble regression to infer the direction and interaction (activation or regression) from regulatory to target genes.

Results

Compared to traditional algorithms, CSER can construct directed networks and infer the type of regulation, thus demonstrating higher accuracy on simulated datasets. Here, using real gene expression data, we applied CSER to construct a colorectal cancer GRN and successfully identified several key regulatory genes closely related to colorectal cancer (CRC), including ADAMDEC1, CLDN8, and GNA11.

Discussion

Importantly, by integrating immune cell and microbial data, we revealed the complex interactions between the CRC gene regulatory network and the tumor microenvironment, providing additional new biomarkers and therapeutic targets for the early diagnosis and prognosis of CRC.

Biologically informed NeuralODEs for genome-wide regulatory dynamics

BACKGROUND

Gene regulatory network (GRN) models that are formulated as ordinary differential equations (ODEs) can accurately explain temporal gene expression patterns and promise to yield new insights into important cellular processes, disease progression, and intervention design. Learning such gene regulatory ODEs is challenging, since we want to predict the evolution of gene expression in a way that accurately encodes the underlying GRN governing the dynamics and the nonlinear functional relationships between genes. Most widely used ODE estimation methods either impose too many parametric restrictions or are not guided by meaningful biological insights, both of which impede either scalability, explainability, or both.

RESULTS

We developed PHOENIX, a modeling framework based on neural ordinary differential equations (NeuralODEs) and Hill-Langmuir kinetics, that overcomes limitations of other methods by flexibly incorporating prior domain knowledge and biological constraints to promote sparse, biologically interpretable representations of GRN ODEs. We tested the accuracy of PHOENIX in a series of in silico experiments, benchmarking it against several currently used tools. We demonstrated PHOENIX's flexibility by modeling regulation of oscillating expression profiles obtained from synchronized yeast cells. We also assessed the scalability of PHOENIX by modeling genome-scale GRNs for breast cancer samples ordered in pseudotime and for B cells treated with Rituximab.

CONCLUSIONS

PHOENIX uses a combination of user-defined prior knowledge and functional forms from systems biology to encode biological "first principles" as soft constraints on the GRN allowing us to predict subsequent gene expression patterns in a biologically explainable manner.

Biologically informed NeuralODEs for genome-wide regulatory dynamics

Modeling dynamics of gene regulatory networks using ordinary differential equations (ODEs) allow a deeper understanding of disease progression and response to therapy, thus aiding in intervention optimization. Although there exist methods to infer regulatory ODEs, these are generally limited to small networks, rely on dimensional reduction, or impose non-biological parametric restrictions - all impeding scalability and explainability. PHOENIX is a neural ODE framework incorporating prior domain knowledge as soft constraints to infer sparse, biologically interpretable dynamics. Extensive experiments - on simulated and real data - demonstrate PHOENIX's unique ability to learn key regulatory dynamics while scaling to the whole genome.

Multiscale modeling of drug resistance in glioblastoma with gene mutations and angiogenesis

Drug resistance is a prominent impediment to the efficacy of targeted therapies across various cancer types, including glioblastoma (GBM). However, comprehending the intricate intracellular and extracellular mechanisms underlying drug resistance remains elusive. Empirical investigations have elucidated that genetic aberrations, such as gene mutations, along with microenvironmental adaptation, notably angiogenesis, act as pivotal drivers of tumor progression and drug resistance. Nonetheless, mathematical models frequently compartmentalize these factors in isolation. In this study, we present a multiscale agent-based model of GBM, encompassing cellular dynamics, intricate signaling pathways, gene mutations, angiogenesis, and therapeutic interventions. This integrative framework facilitates an exploration of the interplay between genetic mutations and the vascular microenvironment in shaping the dynamic evolution of tumors during treatment with tyrosine kinase inhibitor. Our simulations unveil that mutations influencing the migration and proliferation of tumor cells expedite the emergence of phenotype heterogeneity, thereby exacerbating tumor invasion under both treated and untreated conditions. Moreover, angiogenesis proximate to the tumor fosters a protumoral milieu, augmenting mutation-induced drug resistance by increasing the survival rate of tumor cells. Collectively, our findings underscore the dual roles of intrinsic genetic mutations and extrinsic microenvironmental adaptations in steering tumor growth and drug resistance. Finally, we substantiate our model predictions concerning the impact of gene mutations and angiogenesis on the responsiveness of targeted therapies by integrating single-cell RNA-seq, spatial transcriptomics, bulk RNA-seq, and clinical data from GBM patients. The multidimensional approach enhances our understanding of the complexities governing drug resistance in glioma and offers insights into potential therapeutic strategies.

From time-series transcriptomics to gene regulatory networks: A review on inference methods

Inference of gene regulatory networks has been an active area of research for around 20 years, leading to the development of sophisticated inference algorithms based on a variety of assumptions and approaches. With the ever increasing demand for more accurate and powerful models, the inference problem remains of broad scientific interest. The abstract representation of biological systems through gene regulatory networks represents a powerful method to study such systems, encoding different amounts and types of information. In this review, we summarize the different types of inference algorithms specifically based on time-series transcriptomics, giving an overview of the main applications of gene regulatory networks in computational biology. This review is intended to give an updated reference of regulatory networks inference tools to biologists and researchers new to the topic and guide them in selecting the appropriate inference method that best fits their questions, aims, and experimental data.

Integrative bioinformatics analysis of ACS enzymes as candidate prognostic and diagnostic biomarkers in colon adenocarcinoma

Background and purpose

Acyl-CoA synthetase (ACS) enzymes play an important role in the activation of fatty acids. While many studies have found correlations between the expression levels of ACS enzymes with the progression, growth, and survival of cancer cells, their role and expression patterns in colon adenocarcinoma are still greatly unknown and demand further investigation.

Experimental approach

The expression data of colon adenocarcinoma samples were downloaded from the Cancer Genome Atlas (TCGA) database. Normalization and differential expression analysis were performed to identify differentially expressed genes (DEGs). Gene set enrichment analysis was applied to identify top enriched genes from ACS enzymes in cancer samples. Gene ontology and protein-protein interaction analyses were performed for the prediction of molecular functions and interactions. Survival analysis and receiver operating characteristic test (ROC) were performed to find potential prognostic and diagnostic biomarkers.

Findings/Results

ACSL6 and ACSM5 genes demonstrated more significant differential expression and LogFC value compared to other ACS enzymes and also achieved the highest enrichment scores. Gene ontology analysis predicted the involvement of top DEGs in fatty acids metabolism, while protein-protein interaction network analysis presented strong interactions between ACSLs, ACSSs, ACSMs, and ACSBG enzymes with each other. Survival analysis suggested ACSM3 and ACSM5 as potential prognostic biomarkers, while the ROC test predicted stronger diagnostic potential for ACSM5, ACSS2, and ACSF2 genes.

Conclusion and implications

Our findings revealed the expression patterns, prognostic, and diagnostic biomarker potential of ACS enzymes in colon adenocarcinoma. ACSM3, ACSM5, ACSS2, and ACSF2 genes are suggested as possible prognostic and diagnostic biomarkers.

PCB: A pseudotemporal causality-based Bayesian approach to identify EMT-associated regulatory relationships of AS events and RBPs during breast cancer progression

During breast cancer metastasis, the developmental process epithelial-mesenchymal (EM) transition is abnormally activated. Transcriptional regulatory networks controlling EM transition are well-studied; however, alternative RNA splicing also plays a critical regulatory role during this process. Alternative splicing was proved to control the EM transition process, and RNA-binding proteins were determined to regulate alternative splicing. A comprehensive understanding of alternative splicing and the RNA-binding proteins that regulate it during EM transition and their dynamic impact on breast cancer remains largely unknown. To accurately study the dynamic regulatory relationships, time-series data of the EM transition process are essential. However, only cross-sectional data of epithelial and mesenchymal specimens are available. Therefore, we developed a pseudotemporal causality-based Bayesian (PCB) approach to infer the dynamic regulatory relationships between alternative splicing events and RNA-binding proteins. Our study sheds light on facilitating the regulatory network-based approach to identify key RNA-binding proteins or target alternative splicing events for the diagnosis or treatment of cancers. The data and code for PCB are available at: http://hkumath.hku.hk/~wkc/PCB(data+code).zip.

Biologically informed NeuralODEs for genome-wide regulatory dynamics

Models that are formulated as ordinary differential equations (ODEs) can accurately explain temporal gene expression patterns and promise to yield new insights into important cellular processes, disease progression, and intervention design. Learning such ODEs is challenging, since we want to predict the evolution of gene expression in a way that accurately encodes the causal gene-regulatory network (GRN) governing the dynamics and the nonlinear functional relationships between genes. Most widely used ODE estimation methods either impose too many parametric restrictions or are not guided by meaningful biological insights, both of which impedes scalability and/or explainability. To overcome these limitations, we developed PHOENIX, a modeling framework based on neural ordinary differential equations (NeuralODEs) and Hill-Langmuir kinetics, that can flexibly incorporate prior domain knowledge and biological constraints to promote sparse, biologically interpretable representations of ODEs. We test accuracy of PHOENIX in a series of in silico experiments benchmarking it against several currently used tools for ODE estimation. We also demonstrate PHOENIX's flexibility by studying oscillating expression data from synchronized yeast cells and assess its scalability by modelling genome-scale breast cancer expression for samples ordered in pseudotime. Finally, we show how the combination of user-defined prior knowledge and functional forms from systems biology allows PHOENIX to encode key properties of the underlying GRN, and subsequently predict expression patterns in a biologically explainable way.

Timed hazard networks: Incorporating temporal difference for oncogenetic analysis

Oncogenetic graphical models are crucial for understanding cancer progression by analyzing the accumulation of genetic events. These models are used to identify statistical dependencies and temporal order of genetic events, which helps design targeted therapies. However, existing algorithms do not account for temporal differences between samples in oncogenetic analysis. This paper introduces Timed Hazard Networks (TimedHN), a new statistical model that uses temporal differences to improve accuracy and reliability. TimedHN models the accumulation process as a continuous-time Markov chain and includes an efficient gradient computation algorithm for optimization. Our simulation experiments demonstrate that TimedHN outperforms current state-of-the-art graph reconstruction methods. We also compare TimedHN with existing methods on a luminal breast cancer dataset, highlighting its potential utility. The Matlab implementation and data are available at https://github.com/puar-playground/TimedHN.

Dynamic cancer drivers: a causal approach for cancer driver discovery based on bio-pathological trajectories

The traditional way for discovering genes which drive cancer (namely cancer drivers) neglects the dynamic information of cancer development, even though it is well known that cancer progresses dynamically. To enhance cancer driver discovery, we expand cancer driver concept to dynamic cancer driver as a gene driving one or more bio-pathological transitions during cancer progression. Our method refers to the fact that cancer should not be considered as a single process but a compendium of altered biological processes causing the disease to develop over time. Reciprocally, different drivers of cancer can potentially be discovered by analysing different bio-pathological pathways. We propose a novel approach for causal inference of genes driving one or more core processes during cancer development (i.e. dynamic cancer driver). We use the concept of pseudotime for inferring the latent progression of samples along a biological transition during cancer and identifying a critical event when such a process is significantly deviated from normal to carcinogenic. We infer driver genes by assessing the causal effect they have on the process after such a critical event. We have applied our method to single-cell and bulk sequencing datasets of breast cancer. The evaluation results show that our method outperforms well-recognized cancer driver inference methods. These results suggest that including information of the underlying dynamics of cancer improves the inference process (in comparison with using static data), and allows us to discover different sets of driver genes from different processes in cancer. R scripts and datasets can be found at https://github.com/AndresMCB/DynamicCancerDriver.

An Analysis of Transcriptomic Burden Identifies Biological Progression Roadmaps for Hematological Malignancies and Solid Tumors

Biological paths of tumor progression are difficult to predict without time-series data. Using median shift and abacus transformation in the analysis of RNA sequencing data sets, natural patient stratifications were found based on their transcriptomic burden (TcB). Using gene-behavior analysis, TcB groups were evaluated further to discover biological courses of tumor progression. We found that solid tumors and hematological malignancies (n = 4179) share conserved biological patterns, and biological network complexity decreases at increasing TcB levels. An analysis of gene expression datasets including pediatric leukemia patients revealed TcB patterns with biological directionality and survival implications. A prospective interventional study with PI3K targeted therapy in canine lymphomas proved that directional biological responses are dynamic. To conclude, TcB-enriched biological mechanisms detected the existence of biological trajectories within tumors. Using this prognostic informative novel informatics method, which can be applied to tumor transcriptomes and progressive diseases inspires the design of progression-specific therapeutic approaches.

Inferring disease progression and gene regulatory networks from clinical transcriptomic data using PROB_R

Due to a lack of explicit temporal information, it can be challenging to infer gene regulatory networks from clinical transcriptomic data. Here, we describe the protocol of PROB_R for inferring latent temporal disease progression and reconstructing gene regulatory networks from cross-sectional clinical transcriptomic data. We illustrate the protocol by applying it to a breast cancer dataset to demonstrate its use in recovering pseudo-temporal dynamics of gene expression alongside disease progression, reconstructing gene regulatory networks, and identifying key regulatory genes.

For complete details on the use and execution of this protocol, please refer to Sun et al. (2021).

•

Perform downstream analysis of cross-sectional transcriptomic data

•

Infer temporal progression of disease with a graph-based random walk protocol

•

Infer causal gene regulatory networks with an ODE-Bayesian-Lasso protocol

•

Identify key genes from the reconstructed network for further analysis

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Acetyl-CoA Synthetase 2 as a Therapeutic Target in Tumor Metabolism

Simple Summary

Acetyl-CoA Synthetase 2 (ACSS2) is highly expressed in a variety of tumors, which is very important for tumor growth, proliferation, invasion, and metastasis in the nutritional stress microenvironment. Studies have proven that ACSS2 inhibitors can be effective in halting cancer growth and can be combined with other antineoplastic drugs to reduce drug resistance. This article mainly reviews the mechanism of ACSS2-promoting tumor growth from many aspects and the prospect of clinical application of targeted inhibitors.

Abstract

Acetyl-CoA Synthetase 2 (ACSS2) belongs to a member of the acyl-CoA short-chain synthase family, which can convert acetate in the cytoplasm and nucleus into acetyl-CoA. It has been proven that ACSS2 is highly expressed in glioblastoma, breast cancer, liver cancer, prostate cancer, bladder cancer, renal cancer, and other tumors, and is closely related to tumor stage and the overall survival rate of patients. Accumulating studies show that hypoxia and a low serum level induce ACSS2 expression to help tumor cells cope with this nutrient-poor environment. The potential mechanisms are associated with the ability of ACSS2 to promote the synthesis of lipids in the cytoplasm, induce the acetylation of histones in the nucleus, and facilitate the expression of autophagy genes. Novel-specific inhibitors of ACSS2 are developed and confirmed to the effectiveness in pre-clinical tumor models. Targeting ACSS2 may provide novel approaches for tumor treatment. This review summarizes the biological function of ACSS2, its relation to survival and prognosis in different tumors, and how ACSS2 mediates different pathways to promote tumor metastasis, invasion, and drug resistance.