Multiomics with Evolutionary Computation to Identify Molecular and Module Biomarkers for Early Diagnosis and Treatment of Complex Disease

It is important to identify disease biomarkers (DBs) for early diagnosis and treatment of complex diseases in personalized medicine. However, existing methods integrating intelligence technologies and multiomics to predict key biomarkers are limited by the complex dynamic characteristics of omics data, making it difficult to meet the high-precision requirements for biomarker characterization in large dimensions. This study reviewed current analysis methods of evolutionary computation (EC) by considering the essential characteristics of DB identification problems and the advantages of EC, aiming to explore the complex dynamic characteristics of multiomics. In this study, EC-based biomarker identification strategies were summarized as evolutionary algorithms, swarm intelligence and other EC methods for molecular and module DB identification, respectively. Finally, we pointed out the challenges in current research and future research directions. This study can enrich the application of EC theory and promote interdisciplinary integration between EC and bioinformatics.

EIEPCF: accurate inference of functional gene regulatory networks by eliminating indirect effects from confounding factors

Reconstructing functional gene regulatory networks (GRNs) is a primary prerequisite for understanding pathogenic mechanisms and curing diseases in animals, and it also provides an important foundation for cultivating vegetable and fruit varieties that are resistant to diseases and corrosion in plants. Many computational methods have been developed to infer GRNs, but most of the regulatory relationships between genes obtained by these methods are biased. Eliminating indirect effects in GRNs remains a significant challenge for researchers. In this work, we propose a novel approach for inferring functional GRNs, named EIEPCF (eliminating indirect effects produced by confounding factors), which eliminates indirect effects caused by confounding factors. This method eliminates the influence of confounding factors on regulatory factors and target genes by measuring the similarity between their residuals. The validation results of the EIEPCF method on simulation studies, the gold-standard networks provided by the DREAM3 Challenge and the real gene networks of Escherichia coli demonstrate that it achieves significantly higher accuracy compared to other popular computational methods for inferring GRNs. As a case study, we utilized the EIEPCF method to reconstruct the cold-resistant specific GRN from gene expression data of cold-resistant in Arabidopsis thaliana. The source code and data are available at https://github.com/zhanglab-wbgcas/EIEPCF.

An evolutionary learning-based method for identifying a circulating miRNA signature for breast cancer diagnosis prediction

Breast cancer (BC) is one of the most commonly diagnosed cancers worldwide. As key regulatory molecules in several biological processes, microRNAs (miRNAs) are potential biomarkers for cancer. Understanding the miRNA markers that can detect BC may improve survival rates and develop new targeted therapeutic strategies. To identify a circulating miRNA signature for diagnostic prediction in patients with BC, we developed an evolutionary learning-based method called BSig. BSig established a compact set of miRNAs as potential markers from 1280 patients with BC and 2686 healthy controls retrieved from the serum miRNA expression profiles for the diagnostic prediction. BSig demonstrated outstanding prediction performance, with an independent test accuracy and area under the receiver operating characteristic curve were 99.90% and 0.99, respectively. We identified 12 miRNAs, including hsa-miR-3185, hsa-miR-3648, hsa-miR-4530, hsa-miR-4763-5p, hsa-miR-5100, hsa-miR-5698, hsa-miR-6124, hsa-miR-6768-5p, hsa-miR-6800-5p, hsa-miR-6807-5p, hsa-miR-642a-3p, and hsa-miR-6836-3p, which significantly contributed towards diagnostic prediction in BC. Moreover, through bioinformatics analysis, this study identified 65 miRNA-target genes specific to BC cell lines. A comprehensive gene-set enrichment analysis was also performed to understand the underlying mechanisms of these target genes. BSig, a tool capable of BC detection and facilitating therapeutic selection, is publicly available at https://github.com/mingjutsai/BSig.

Developing a practical neurodevelopmental prediction model for targeting high-risk very preterm infants during visit after NICU: a retrospective national longitudinal cohort study

BACKGROUND

Follow-up visits for very preterm infants (VPI) after hospital discharge is crucial for their neurodevelopmental trajectories, but ensuring their attendance before 12 months corrected age (CA) remains a challenge. Current prediction models focus on future outcomes at discharge, but post-discharge data may enhance predictions of neurodevelopmental trajectories due to brain plasticity. Few studies in this field have utilized machine learning models to achieve this potential benefit with transparency, explainability, and transportability.

METHODS

We developed four prediction models for cognitive or motor function at 24 months CA separately at each follow-up visits, two for the 6-month and two for the 12-month CA visits, using hospitalized and follow-up data of VPI from the Taiwan Premature Infant Follow-up Network from 2010 to 2017. Regression models were employed at 6 months CA, defined as a decline in The Bayley Scales of Infant Development 3rd edition (BSIDIII) composite score > 1 SD between 6- and 24-month CA. The delay models were developed at 12 months CA, defined as a BSIDIII composite score < 85 at 24 months CA. We used an evolutionary-derived machine learning method (EL-NDI) to develop models and compared them to those built by lasso regression, random forest, and support vector machine.

RESULTS

One thousand two hundred forty-four VPI were in the developmental set and the two validation cohorts had 763 and 1347 VPI, respectively. EL-NDI used only 4-10 variables, while the others required 29 or more variables to achieve similar performance. For models at 6 months CA, the area under the receiver operating curve (AUC) of EL-NDI were 0.76-0.81(95% CI, 0.73-0.83) for cognitive regress with 4 variables and 0.79-0.83 (95% CI, 0.76-0.86) for motor regress with 4 variables. For models at 12 months CA, the AUC of EL-NDI were 0.75-0.78 (95% CI, 0.72-0.82) for cognitive delay with 10 variables and 0.73-0.82 (95% CI, 0.72-0.85) for motor delay with 4 variables.

CONCLUSIONS

Our EL-NDI demonstrated good performance using simpler, transparent, explainable models for clinical purpose. Implementing these models for VPI during follow-up visits may facilitate more informed discussions between parents and physicians and identify high-risk infants more effectively for early intervention.

GMFGRN: a matrix factorization and graph neural network approach for gene regulatory network inference

The recent advances of single-cell RNA sequencing (scRNA-seq) have enabled reliable profiling of gene expression at the single-cell level, providing opportunities for accurate inference of gene regulatory networks (GRNs) on scRNA-seq data. Most methods for inferring GRNs suffer from the inability to eliminate transitive interactions or necessitate expensive computational resources. To address these, we present a novel method, termed GMFGRN, for accurate graph neural network (GNN)-based GRN inference from scRNA-seq data. GMFGRN employs GNN for matrix factorization and learns representative embeddings for genes. For transcription factor-gene pairs, it utilizes the learned embeddings to determine whether they interact with each other. The extensive suite of benchmarking experiments encompassing eight static scRNA-seq datasets alongside several state-of-the-art methods demonstrated mean improvements of 1.9 and 2.5% over the runner-up in area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC). In addition, across four time-series datasets, maximum enhancements of 2.4 and 1.3% in AUROC and AUPRC were observed in comparison to the runner-up. Moreover, GMFGRN requires significantly less training time and memory consumption, with time and memory consumed <10% compared to the second-best method. These findings underscore the substantial potential of GMFGRN in the inference of GRNs. It is publicly available at https://github.com/Lishuoyy/GMFGRN.

iLSGRN: inference of large-scale gene regulatory networks based on multi-model fusion

MOTIVATION

Gene regulatory networks (GRNs) are a way of describing the interaction between genes, which contribute to revealing the different biological mechanisms in the cell. Reconstructing GRNs based on gene expression data has been a central computational problem in systems biology. However, due to the high dimensionality and non-linearity of large-scale GRNs, accurately and efficiently inferring GRNs is still a challenging task.

RESULTS

In this article, we propose a new approach, iLSGRN, to reconstruct large-scale GRNs from steady-state and time-series gene expression data based on non-linear ordinary differential equations. Firstly, the regulatory gene recognition algorithm calculates the Maximal Information Coefficient between genes and excludes redundant regulatory relationships to achieve dimensionality reduction. Then, the feature fusion algorithm constructs a model leveraging the feature importance derived from XGBoost (eXtreme Gradient Boosting) and RF (Random Forest) models, which can effectively train the non-linear ordinary differential equations model of GRNs and improve the accuracy and stability of the inference algorithm. The extensive experiments on different scale datasets show that our method makes sensible improvement compared with the state-of-the-art methods. Furthermore, we perform cross-validation experiments on the real gene datasets to validate the robustness and effectiveness of the proposed method.

AVAILABILITY AND IMPLEMENTATION

The proposed method is written in the Python language, and is available at: https://github.com/lab319/iLSGRN.

Prediction of extranodal extension in head and neck squamous cell carcinoma by CT images using an evolutionary learning model

BACKGROUND

Extranodal extension (ENE) in head and neck squamous cell carcinoma (HNSCC) correlates to poor prognoses and influences treatment strategies. Deep learning may yield promising performance of predicting ENE in HNSCC but lack of transparency and interpretability. This work proposes an evolutionary learning method, called EL-ENE, to establish a more interpretable ENE prediction model for aiding clinical diagnosis.

METHODS

There were 364 HNSCC patients who underwent neck lymph node (LN) dissection with pre-operative contrast-enhanced computerized tomography images. All the 778 LNs were divided into training and test sets with the ratio 8:2. EL-ENE uses an inheritable bi-objective combinatorial genetic algorithm for optimal feature selection and parameter setting of support vector machine. The diagnostic performances of the ENE prediction model and radiologists were compared using independent test datasets.

RESULTS

The EL-ENE model achieved the test accuracy of 80.00%, sensitivity of 81.13%, and specificity of 79.44% for ENE detection. The three radiologists achieved the mean diagnostic accuracy of 70.4%, sensitivity of 75.6%, and specificity of 67.9%. The features of gray-level texture and 3D morphology of LNs played essential roles in predicting ENE.

CONCLUSIONS

The EL-ENE method provided an accurate, comprehensible, and robust model to predict ENE in HNSCC with interpretable radiomic features for expanding clinical knowledge. The proposed transparent prediction models are more trustworthy and may increase their acceptance in daily clinical practice.

Artificial intelligence-driven pan-cancer analysis reveals miRNA signatures for cancer stage prediction

The ability to detect cancer at an early stage in patients who would benefit from effective therapy is a key factor in increasing survivability. This work proposes an evolutionary supervised learning method called CancerSig to identify cancer stage-specific microRNA (miRNA) signatures for early cancer predictions. CancerSig established a compact panel of miRNA signatures as potential markers from 4,667 patients with 15 different types of cancers for the cancer stage prediction, and achieved a mean performance: 10-fold cross-validation accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve of 84.27% ± 6.31%, 0.81 ± 0.12, 0.80 ± 0.10, and 0.80 ± 0.06, respectively. The pan-cancer analysis of miRNA signatures suggested that three miRNAs, hsa-let-7i-3p, hsa-miR-362-3p, and hsa-miR-3651, contributed significantly toward stage prediction across 8 cancers, and each of the 67 miRNAs of the panel was a biomarker of stage prediction in more than one cancer. CancerSig may serve as the basis for cancer screening and therapeutic selection..

Toward surface defect detection in electronics manufacturing by an accurate and lightweight YOLO-style object detector

In electronics manufacturing, surface defect detection is very important for product quality control, and defective products can cause severe customer complaints. At the same time, in the manufacturing process, the cycle time of each product is usually very short. Furthermore, high-resolution input images from high-resolution industrial cameras are necessary to meet the requirements for high quality control standards. Hence, how to design an accurate object detector with real-time inference speed that can accept high-resolution input is an important task. In this work, an accurate YOLO-style object detector was designed, ATT-YOLO, which uses only one self-attention module, many-scale feature extraction and integration in the backbone and feature pyramid, and an improved auto-anchor design to address this problem. There are few datasets for surface detection in electronics manufacturing. Hence, we curated a dataset consisting of 14,478 laptop surface defects, on which ATT-YOLO achieved 92.8% mAP0.5 for the binary-class object detection task. We also further verified our design on the COCO benchmark dataset. Considering both computation costs and the performance of object detectors, ATT-YOLO outperforms several state-of-the-art and lightweight object detectors on the COCO dataset. It achieves a 44.9% mAP score and 21.8 GFLOPs, which is better than the compared models including YOLOv8-small (44.9%, 28.6G), YOLOv7-tiny-SiLU (38.7%, 13.8G), YOLOv6-small (43.1%, 44.2G), pp-YOLOE-small (42.7%, 17.4G), YOLOX-small (39.6%, 26.8G), and YOLOv5-small (36.7%, 17.2G). We hope that this work can serve as a useful reference for the utilization of attention-based networks in real-world situations.

Combining kinetic orders for efficient S-System modelling of gene regulatory network

S-System models, non-linear differential equation models, are widely used for reconstructing gene regulatory networks from temporal gene expression data. An S-System model involves two states, generation and degeneration, and uses the kinetic parameters g_ij and h_ij, to represent the direction, nature, and intensity of the genetic interactions. The need for learning a large number of model parameters results in increased computational expense. Previously, we improved the performance of the algorithm using dynamic allocation of the maximum in-degree for each gene. While the method was effective for smaller networks, a large amount of computation was still needed for larger networks. This problem arose mainly due to the increased occurrence of invalid networks during optimization, primarily because the two kinetic parameters (g_ij and h_ij) of the S-System model converge independently during optimization. Being independent, these two parameters can converge to values that can indicate contradictory gene interactions, specifically inhibition or activation. In this study, to address this major challenge in S-System modelling, we developed a novel method that includes two features: a penalty term that penalizes those networks with invalid kinetic orders, and a parameter, w_ij, derived by combining the kinetic parameters g_ij and h_ij. The novel penalty term was used for candidate selection during the process of optimizing the DRNI (Dynamically Regulated Network Initialization) algorithm. Rather than remaining constant, it is dynamic, with its magnitude dependent on the number of invalid interactions in the given network. This approach encourages the generation of valid candidate solutions, and eliminates invalid networks in a systematic manner. The previous DRNI method, a two-stage approach which uses dynamic allocation of the maximum in-degree for each gene, was further improved by adding a third stage which applies the proposed w_ij to handle the invalid regulations that may still exist in that candidate solutions. The method was tested on different gene expression datasets, and was able to reduce the number of iterations and produce improved network accuracies. For a 20 gene network, the number of generations required for convergence was reduced by 300, and the F-score improved by 0.05 compared to our previously reported DRNI approach. For the well-known 10 gene networks of the DREAM challenge, our method produced an improvement in the average area under the ROC curve of the DREAM4 10 gene networks.

Tracking the amino acid changes of spike proteins across diverse host species of severe acute respiratory syndrome coronavirus 2

Knowledge of the host-specific properties of the spike protein is of crucial importance to understand the adaptability of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) to infect multiple species and alter transmissibility, particularly in humans. Here, we propose a spike protein predictor SPIKES incorporating with an inheritable bi-objective combinatorial genetic algorithm to identify the biochemical properties of spike proteins and determine their specificity to human hosts. SPIKES identified 20 informative physicochemical properties of the spike protein, including information measures for alpha helix and relative mutability, and amino acid and dipeptide compositions, which have shown compositional difference at the amino acid sequence level between human and diverse animal coronaviruses. We suggest that alterations of these amino acids between human and animal coronaviruses may provide insights into the development and transmission of SARS-CoV-2 in human and other species and support the discovery of targeted antiviral therapies.

•

Differences exist in the amino acids within the S protein of diverse host species CoVs

•

We developed SPIKES to identify informative properties of S protein

•

SARS-CoV-2 variants have amino acid changes that alter infection and transmission

•

The SPIKES identified changes in S protein properties from animal to human host CoVs

Inference of genetic regulatory networks with regulatory hubs using vector autoregressions and automatic relevance determination with model selections

The inference of genetic regulatory networks (GRNs) reveals how genes interact with each other. A few genes can regulate many genes as targets to control cell functions. We present new methods based on the order-1 vector autoregression (VAR1) for inferring GRNs from gene expression time series. The methods use the automatic relevance determination (ARD) to incorporate the regulatory hub structure into the estimation of VAR1 in a Bayesian framework. Several sparse approximation schemes are applied to the estimated regression weights or VAR1 model to generate the sparse weighted adjacency matrices representing the inferred GRNs. We apply the proposed and several widespread reference methods to infer GRNs with up to 100 genes using simulated, DREAM4 in silico and experimental E. coli gene expression time series. We show that the proposed methods are efficient on simulated hub GRNs and scale-free GRNs using short time series simulated by VAR1s and outperform reference methods on small-scale DREAM4 in silico GRNs and E. coli GRNs. They can utilize the known major regulatory hubs to improve the performance on larger DREAM4 in silico GRNs and E. coli GRNs. The impact of nonlinear time series data on the performance of proposed methods is discussed.

Evolutionary Learning-Derived Clinical-Radiomic Models for Predicting Early Recurrence of Hepatocellular Carcinoma after Resection

BACKGROUND AND AIMS

Current prediction models for early recurrence of hepatocellular carcinoma (HCC) after surgical resection remain unsatisfactory. The aim of this study was to develop evolutionary learning-derived prediction models with interpretability using both clinical and radiomic features to predict early recurrence of HCC after surgical resection.

METHODS

Consecutive 517 HCC patients receiving surgical resection with available contrast-enhanced computed tomography (CECT) images before resection were retrospectively enrolled. Patients were randomly assigned to a training set (n = 362) and a test set (n = 155) in a ratio of 7:3. Tumor segmentation of all CECT images including noncontrast phase, arterial phase, and portal venous phase was manually performed for radiomic feature extraction. A novel evolutionary learning-derived method called genetic algorithm for predicting recurrence after surgery of liver cancer (GARSL) was proposed to design prediction models for early recurrence of HCC within 2 years after surgery.

RESULTS

A total of 143 features, including 26 preoperative clinical features, 5 postoperative pathological features, and 112 radiomic features were used to develop GARSL preoperative and postoperative models. The area under the receiver operating characteristic curves (AUCs) for early recurrence of HCC within 2 years were 0.781 and 0.767, respectively, in the training set, and 0.739 and 0.741, respectively, in the test set. The accuracy of GARSL models derived from the evolutionary learning method was significantly better than models derived from other well-known machine learning methods or the early recurrence after surgery for liver tumor (ERASL) preoperative (AUC = 0.687, p < 0.001 vs. GARSL preoperative) and ERASL postoperative (AUC = 0.688, p < 0.001 vs. GARSL postoperative) models using clinical features only.

CONCLUSION

The GARSL models using both clinical and radiomic features significantly improved the accuracy to predict early recurrence of HCC after surgical resection, which was significantly better than other well-known machine learning-derived models and currently available clinical models.

Identification and Characterization of Species-Specific Severe Acute Respiratory Syndrome Coronavirus 2 Physicochemical Properties

There is an urgent need to elucidate the underlying mechanisms of coronavirus disease (COVID-19) so that vaccines and treatments can be devised. Severe acute respiratory syndrome coronavirus 2 has genetic similarity with bats and pangolin viruses, but a comprehensive understanding of the functions of its proteins at the amino acid sequence level is lacking. A total of 4320 sequences of human and nonhuman coronaviruses was retrieved from the Global Initiative on Sharing All Influenza Data and the National Center for Biotechnology Information. This work proposes an optimization method COVID-Pred with an efficient feature selection algorithm to classify the species-specific coronaviruses based on physicochemical properties (PCPs) of their sequences. COVID-Pred identified a set of 11 PCPs using a support vector machine and achieved 10-fold cross-validation and test accuracies of 99.53% and 97.80%, respectively. These findings could provide key insights into understanding the driving forces during the course of infection and assist in developing effective therapies.

A comprehensive overview and critical evaluation of gene regulatory network inference technologies

Gene regulatory network (GRN) is the important mechanism of maintaining life process, controlling biochemical reaction and regulating compound level, which plays an important role in various organisms and systems. Reconstructing GRN can help us to understand the molecular mechanism of organisms and to reveal the essential rules of a large number of biological processes and reactions in organisms. Various outstanding network reconstruction algorithms use specific assumptions that affect prediction accuracy, in order to deal with the uncertainty of processing. In order to study why a certain method is more suitable for specific research problem or experimental data, we conduct research from model-based, information-based and machine learning-based method classifications. There are obviously different types of computational tools that can be generated to distinguish GRNs. Furthermore, we discuss several classical, representative and latest methods in each category to analyze core ideas, general steps, characteristics, etc. We compare the performance of state-of-the-art GRN reconstruction technologies on simulated networks and real networks under different scaling conditions. Through standardized performance metrics and common benchmarks, we quantitatively evaluate the stability of various methods and the sensitivity of the same algorithm applying to different scaling networks. The aim of this study is to explore the most appropriate method for a specific GRN, which helps biologists and medical scientists in discovering potential drug targets and identifying cancer biomarkers.

Inferring dynamic gene regulatory networks with low-order conditional independencies – an evaluation of the method

Over a decade ago, Lèbre (2009) proposed an inference method, G1DBN, to learn the structure of gene regulatory networks (GRNs) from high dimensional, sparse time-series gene expression data. Their approach is based on concept of low-order conditional independence graphs that they extend to dynamic Bayesian networks (DBNs). They present results to demonstrate that their method yields better structural accuracy compared to the related Lasso and Shrinkage methods, particularly where the data is sparse, that is, the number of time measurements n is much smaller than the number of genes p. This paper challenges these claims using a careful experimental analysis, to show that the GRNs reverse engineered from time-series data using the G1DBN approach are less accurate than claimed by Lèbre (2009). We also show that the Lasso method yields higher structural accuracy for graphs learned from the simulated data, compared to the G1DBN method, particularly when the data is sparse ($n{< }{< }p$). The Lasso method is also better than G1DBN at identifying the transcription factors (TFs) involved in the cell cycle of Saccharomyces cerevisiae.