Sequence Learning Research Articles

Pooled analysis of oral microbiome profiles defines robust signatures associated with periodontitis.

Oral microbial dysbiosis has been associated with periodontitis in studies using 16S rRNA gene sequencing analysis. However, this technology is not sufficient to consistently separate the bacterial species to species level, and reproducible oral microbiome signatures are scarce. Here, using ultra-deep metagenomic sequencing and machine learning tools, we defined a simple two-factor decision tree, based on Tannerella forsythia and Fretibacterium fastidiosum, that was highly associated with periodontitis. Altogether, we defined robust oral microbiome signatures relevant to the pathophysiology of periodontitis that can help define promising targets for microbiome therapeutic modulation when caring for patients with periodontitis.

Motor sequence-specific learning-induced corticospinal plasticity is biased towards sensorimotor mu rhythm peak phases.

Motor cortical (M1) transcranial magnetic stimulation (TMS) interventions increase corticospinal output and improve motor learning when delivered during sensorimotor mu rhythm trough but not peak phases, suggesting that the mechanisms supporting motor learning may be most active during mu trough phases. Based on these findings, we predicted that motor sequence learning-related corticospinal plasticity would be most evident when measured during mu trough phases. Healthy adults were assigned to either a sequence or random group. Participants in the sequence group practiced the implicit serial reaction time task (SRTT), which contained an embedded, repeating 12-item sequence. Participants in the random group practiced a version of the SRTT that contained no sequence. We measured mu phase-independent and mu phase-dependent MEP amplitudes using EEG-informed single-pulse TMS before, immediately, and 30 minutes after the SRTT in both groups. All participants performed a retention test one hour after SRTT acquisition. In both groups, mu phase-independent MEP amplitudes increased following SRTT acquisition, but the pattern of mu phase-dependent MEP amplitude increases after SRTT acquisition differed between groups. Relative to the random group, the sequence group showed greater increases in peak-specific corticospinal output 30 minutes after SRTT acquisition. Contrary to our original hypothesis, results revealed that motor sequence learning recruits peak-rather than trough-specific neurophysiological mechanisms. These findings suggest that mu peak phases may provide protected time windows for motor memory consolidation and demonstrate the presence of a mu phase-dependent motor learning mechanism in the human brain.

Design, Development, Implementation, and Evaluation of a Teaching–Learning Sequence on Heat Conduction

Design, Development, Implementation, and Evaluation of a Teaching–Learning Sequence on Heat Conduction

The importance of coding and translation between programming languages in sequential activities of pre-service teachers: an approach

AbstractIn the last two decades, computational thinking has gained wide relevance in international educational systems. The inclusion of this new type of thinking poses educational challenges with some underlying research questions that need to be answered to meet these challenges with quality. Thus, this study focuses on analyzing the difficulties that teachers in initial training experience have, when carrying out translation tasks of programming languages used by certain educational robots, in this case, the Cubetto. For this purpose, a specific learning sequence has been designed to work with different programming languages (Cubetto, Bee-Bot, Scratch) and natural language. The work of early childhood and elementary trainee teachers in these tasks has been analyzed using a descriptive approach. The main results are: (1) some of the difficulties encountered are clearly caused by the Cubetto hardware (regardless of the language to which it is translated) and (2) the designed learning sequence has enabled coding skills to be improved remarkably. We conclude that translation tasks between programming languages are necessary in initial teacher training to improve their ability programming and their computational thinking, and for them to be able to detect the disadvantages and benefits of educational robots in their transposition to the classroom.

Digital Grading Company: Deep Neural Networks for Predictive Analytics for Alternative Asset Portfolio Management

In this study, we investigate the application of deep neural networks for predictive analytics in the realm of alternative asset portfolio management, focusing on Trading Card Games (TCGs) like Pokemon. Our primary objective is to understand the optimal trading cards for investment and to determine the appropriate timing for divestment and retention of these assets. Drawing on the successful implementation of machine learning predictive analytics in traditional financial markets, we aim to extend these techniques to alternative investments. To achieve this, we employ a two-pronged approach using a comprehensive dataset of historical sales values for Pokemon TCGs spanning five years, comprising over 200 million unlabeled data points. The first approach utilizes a semi-supervised sequence learning method, where a Long Short-Term Memory (LSTM) model is initially trained as a sequence autoencoder. The pre-trained model's weights are then used in a supervised phase to fine-tune the predictions for future trading card values. The second approach involves heuristic pretraining, which leverages market-derived indicators from existing literature to bridge the gap between traditional feature engineering and deep learning methods. Our technology stack is designed to handle end-to-end data collection, analytics generation, and execution of recommended trading strategies. Simulated backtesting results demonstrate that our Digital Grading Company (DGC) portfolio model outperforms traditional indices, achieving a 357% return compared to the 61% return of the S&P 500 Index over the same period. Real-time analysis further indicates that our model consistently identifies profitable trading opportunities in the TCG market using hourly data and proprietary algorithmic trade logic, executing trades 24/7. The DGC Predictive Portfolio, based on ungraded Pokemon TCG trading cards, also shows superior performance when compared to the Nasdaq 100 Index (Invesco QQQ ETF). Our analytics framework not only predicts the direction of price movement but also provides insights into the magnitude of changes, enabling a more dynamic and responsive trading strategy. The findings suggest that the integration of deep learning models with heuristic indicators has the potential to revolutionize portfolio management in alternative investments, paving the way for more accurate and profitable decision-making in niche markets. This research highlights the promising role of machine learning in enhancing alternative asset management strategies, suggesting future applications in other asset classes. We conclude that deep neural networks, when combined with data-driven heuristics, can significantly enhance trading outcomes, making them valuable tools for investors in the evolving landscape of alternative investments.

Role of dopamine in reward expectation and predictability during execution of action sequences.

Reward-associated cues serve different functions depending on whether they precede or terminate action sequences. Cues that precede action sequences and signal opportunity for reward could serve as GO signals to initiate the sequence, whereas sequence termination cues could serve as response feedback by signaling reward delivery. Reward expectation during sequence execution depends on these cues and might condition whether behavior is habitual or goal-directed. However, it remains unknown how sequence initiation and termination cues differentially affect reward expectation and contribute to habit learning. Further, while mesolimbic dopamine plays a key role in cue-induced reward expectation and sequence learning, how dynamic changes in dopamine signals differ depending on the response strategy is unclear. Here, we determined how mesolimbic DA signals change over training during cue-mediated sequence learning, depending on the type of cue and the nature of behavioral control. We found sequence initiation and termination cues differentially affect reward expectation during action sequences, with the termination cue contributing to habit and automaticity. Distinct response strategies induced by sequence initiation and termination cues induced differential changes in mesolimbic DA signals that captured variations in reward expectation along sequence execution. Notably, habit-like behavior induced by the sequence termination cue was associated with a rapid shift in DA signals from reward retrieval to the cue. This habit-like behavior was reflected in behavioral inflexibility and attenuated DA reward prediction error signals. Finally, using optogenetics, we provide evidence that phasic DA activity elicited by the sequence termination cue is critical for the development of habit-like behavior.

Identification of novel biomarkers for atherosclerosis using single-cell RNA sequencing and machine learning.

Atherosclerosis (AS) is a predominant etiological factor in numerous cardiovascular diseases, with its associated complications such as myocardial infarction and stroke serving as major contributors to worldwide mortality rates. Here, we devised dependable AS-related biomarkers through the utilization of single-cell RNA sequencing, weighted co-expression network (WGCNA), and differential expression analysis. Furthermore, we employed various machine learning techniques (LASSO and SVM-RFE) to enhance the identification of AS biomarkers, subsequently validating them using the GEO dataset. Following this, CIBERSORT was employed to investigate the correlation between biomarkers and infiltrating immune cells. Consequently, 256 differentially expressed genes (DEGs) were selected in samples of AS and normal. GO and KEGG analyses indicated that these DEGs may be related to the negative regulation of leukocyte-mediated immunity, leukocyte cell-cell adhesion, and immune system processes. Notably, C1QC and COL1A1 were pinpointed as potential diagnostic markers for AS, a finding that was further validated in the GSE21545 dataset. Moreover, the area under the curve (AUC) values for these markers exceeded 0.8, underscoring their diagnostic utility. Analysis of immune cell infiltration revealed that the expression of C1QC was correlated with M0 macrophages, gamma delta T cells, activated mast cells and memory B cells. Similarly, COL1A1 expression was linked to M0 macrophages, memory B cells, activated mast cells, gamma delta T cells, and CD4 native T cells. Finally, these results were validated using mice and human samples through immunofluorescence, immunohistochemistry, and ELISA analysis. Overall, C1QC and COL1A1 would be potential biomarkers for AS diagnosis, and that would provides novel perspectives on the diagnosis and treatment of AS.

Integrating whole genome sequencing and machine learning for predicting antimicrobial resistance in critical pathogens: a systematic review of antimicrobial susceptibility tests.

Infections caused by antibiotic-resistant bacteria pose a major challenge to modern healthcare. This systematic review evaluates the efficacy of machine learning (ML) approaches in predicting antimicrobial resistance (AMR) in critical pathogens (CP), considering Whole Genome Sequencing (WGS) and antimicrobial susceptibility testing (AST). The search covered databases including PubMed/MEDLINE, EMBASE, Web of Science, SCOPUS, and SCIELO, from their inception until June 2024. The review protocol was officially registered on PROSPERO (CRD42024543099). The review included 26 papers, analyzing data from 104,141 microbial samples. Random Forest (RF), XGBoost, and logistic regression (LR) emerged as the top-performing models, with mean Area Under the Receiver Operating Characteristic (AUC) values of 0.89, 0.87, and 0.87, respectively. RF showed superior performance with AUC values ranging from 0.66 to 0.97, while XGBoost and LR showed similar performance with AUC values ranging from 0.83 to 0.91 and 0.76 to 0.96, respectively. Most studies indicate that integrating WGS and AST data into ML models enhances predictive performance, improves antibiotic stewardship, and provides valuable clinical decision support. ML shows significant promise for predicting AMR by integrating WGS and AST data in CP. Standardized guidelines are needed to ensure consistency in future research.

Implicit motor sequence learning using three-dimensional reaching movements with the non-dominant left arm.

Interlimb differences in reach control could impact the learning of a motor sequence that requires whole-arm movements. The purpose of this study was to investigate the learning of an implicit, 3-dimensional whole-arm sequence task with the non-dominant left arm compared to the dominant right arm. Thirty-one right-hand dominant adults completed two consecutive days of practice of a motor sequence task presented in a virtual environment with either their dominant right or non-dominant left arm. Targets were presented one-at-a-time alternating between Random and Repeated sequences. Task performance was indicated by the time to complete the sequence (response time), and kinematic measures (hand path distance, peak velocity) were used to examine how movements changed over time. While the Left Arm group was slower than the Right Arm group at baseline, both groups significantly improved response time with practice with the Left Arm group demonstrating greater gains. The Left Arm group improved performance by decreasing hand path distance (straighter path to targets) while the Right Arm group improved performance through a smaller decrease in hand path distance combined with increasing peak velocity. Gains made during practice on Day 1 were retained on Day 2 for both groups. Overall, individuals reaching with the non-dominant left arm learned the whole-arm motor sequence task but did so through a different strategy than individuals reaching with the dominant right arm. The strategy adopted for the learning of movement sequences that require whole-arm movements may be impacted by differences in reach control between the nondominant and dominant arms.

Dynamic landscape of m6A modifications and related post-transcriptional events in muscle-invasive bladder cancer

BackgroundMuscle-invasive bladder carcinoma (MIBC) is a serious and more advanced stage of bladder carcinoma. N6-Methyladenosine (m6A) is a dynamic and reversible modifications that primarily affects RNA stability and alternative splicing. The dysregulation of m6A in MIBC can be potential target for clinical interventions, but there have been limited studies on m6A modifications in MIBC and their associations with post-transcriptional regulatory processes.MethodsPaired tumor and adjacent-normal tissues were obtained from three patients with MIBC following radical cystectomy. The additional paired tissues for validation were obtained from patients underwent transurethral resection. Utilizing Nanopore direct-RNA sequencing, we characterized the m6A RNA methylation landscape in MIBC, with a focus on identifying post-transcriptional events potentially affected by changes in m6A sites. This included an examination of differential transcript usage, polyadenylation signal sites, and variations in poly(A) tail length, providing insights into the broader impact of m6A alterations on RNA processing in MIBC.ResultsThe prognostic-related m6A genes and m6A-risk model constructed by machine learning enables the stratification of high and low-risk patients with precision. A novel m6A modification site in the 3’ untranslated region (3’UTR) of IGLL5 gene were identified, characterized by a lower m6A methylation ratio, elongated poly(A) tails, and a notable bias in transcript usage. Furthermore, we discovered two particular transcripts, VWA1-203 and CEBPB-201. VWA1-203 displayed diminished m6A methylation levels, a truncated 3’UTR, and an elongated poly(A) tail, whereas CEBPB-201 showed opposite trends, highlighting the complex interplay between m6A modifications and RNA processing. Source code was provided on GitHub (https://github.com/lelelililele/Nanopore-m6A-analysis).ConclusionsThe state-of-the-art Nanopore direct-RNA sequencing and machine learning techniques enables comprehensive identification of m6A modification and provided insights into the potential post-transcriptional regulation mechanisms on the development and progression in MIBC.

GeneCompass: deciphering universal gene regulatory mechanisms with a knowledge-informed cross-species foundation model.

Deciphering universal gene regulatory mechanisms in diverse organisms holds great potential for advancing our knowledge of fundamental life processes and facilitating clinical applications. However, the traditional research paradigm primarily focuses on individual model organisms and does not integrate various cell types across species. Recent breakthroughs in single-cell sequencing and deep learning techniques present an unprecedented opportunity to address this challenge. In this study, we built an extensive dataset of over 120 million human and mouse single-cell transcriptomes. After data preprocessing, we obtained 101,768,420 single-cell transcriptomes and developed a knowledge-informed cross-species foundation model, named GeneCompass. During pre-training, GeneCompass effectively integrated four types of prior biological knowledge to enhance our understanding of gene regulatory mechanisms in a self-supervised manner. By fine-tuning for multiple downstream tasks, GeneCompass outperformed state-of-the-art models in diverse applications for a single species and unlocked new realms of cross-species biological investigations. We also employed GeneCompass to search for key factors associated with cell fate transition and showed that the predicted candidate genes could successfully induce the differentiation of human embryonic stem cells into the gonadal fate. Overall, GeneCompass demonstrates the advantages of using artificial intelligence technology to decipher universal gene regulatory mechanisms and shows tremendous potential for accelerating the discovery of critical cell fate regulators and candidate drug targets.

Machine Learning for Genomic Expression Classification-Based Phenotype Prediction in Topological Data Analysis

Genomic data has become more prevalent due to sequencing and Machine Learning (ML) innovations, which have increased the biological genomics study. The multidimensional nature of this data provides challenges to phenotype prediction, which is required for individualized health care and the research investigation of genetic problems; nevertheless, it holds tremendous potential for understanding the association between genes and physical features. The authors of this paper introduce a new technique for symptom prediction from data from genomes, which combines Topological Data Analysis (TDA), Graph Convolutional Networks (GCN), and Support Vector Machines (SVM). The proposed method aims to address these challenges. By using TDA for multifaceted feature extraction, GCN to analyze gene interaction networks, and SVM for reliable classification in high-dimensional spaces, the above technique overcomes the drawbacks of conventional approaches. This TDA-GCN-SVM model has been demonstrated to be implemented in a method that is superior to conventional methods on distinct tumor datasets in terms of accuracy and additional measures. A novel method for genomic study and a more significant comprehension of genomic data analysis are both caused by this innovation, which is an enormous achievement in precision healthcare.

IPSRM: An intent perceived sequential recommendation model

Objectives:Sequential recommendation aims to recommend items that are relevant to users’ interests based on their existing interaction sequences. Current models lack in capturing users’ latent intentions and do not sufficiently consider sequence information during the modeling of users and items. Additionally, noise in user interaction sequences can affect the model’s optimization process. Methods:This paper introduces an intent perceived sequential recommendation model (IPSRM). IPSRM employs the generalized expectation–maximization (EM) framework, alternating between learning sequence representations and optimizing the model to better capture the underlying intentions of user interactions. Specifically, IPSRM maps unlabeled behavioral sequences into frequency domain filtering and random Gaussian distribution space. These mappings reduce the impact of noise and improve the learning of user behavior representations. Through clustering process, IPSRM captures users’ potential interaction intentions and incorporates them as one of the supervisions into the contrastive self-supervised learning process to guide the optimization process. Results:Experimental results on four standard datasets demonstrate the superiority of IPSRM. Comparative experiments also verify that IPSRM exhibits strong robustness under cold start and noisy interaction conditions. Conclusions:Capturing latent user intentions, integrating intention-based supervision into model optimization, and mitigating noise in sequential modeling significantly enhance the performance of sequential recommendation systems.

The role of lactylation in plasma cells and its impact on rheumatoid arthritis pathogenesis: insights from single-cell RNA sequencing and machine learning.

Rheumatoid arthritis (RA) is a chronic autoimmune disorder characterized by persistent synovitis, systemic inflammation, and autoantibody production. This study aims to explore the role of lactylation in plasma cells and its impact on RA pathogenesis. We utilized single-cell RNA sequencing (scRNA-seq) data and applied bioinformatics and machine learning techniques. A total of 10,163 cells were retained for analysis after quality control. Clustering analysis identified 13 cell clusters, with plasma cells displaying the highest lactylation scores. We performed pathway enrichment analysis to examine metabolic activity, such as oxidative phosphorylation and glycolysis, in highly lactylated plasma cells. Additionally, we employed 134 machine learning algorithms to identify seven core lactylation-promoting genes and constructed a diagnostic model with an average AUC of 0.918. The RA lactylation score (RAlac_score) was significantly elevated in RA patients and positively correlated with immune cell infiltration and immune checkpoint molecule expression. Differential expression analysis between two plasma cell clusters revealed distinct metabolic and immunological profiles, with cluster 2 demonstrating increased immune activity and extracellular matrix interactions. qRT-PCR validation confirmed that NDUFB3, NGLY1, and SLC25A4 are highly expressed in RA. This study highlights the critical role of lactylation in plasma cells for RA pathogenesis and identifies potential biomarkers and therapeutic targets, which may offer insights for future therapeutic strategies.

The problem with two-event sequence learning by pigeons

Bonobos appear to show little evidence of learning to make one response (R1) to an AB sequence and a different response (R2) to sequences BB, AA, and BA (Lind et al. PLoS ONE 18(9):e0290546, 2023), yet under different conditions, pigeons can learn this (Weisman et al. Exp Psychol Anim Behav Process 6(4):312, 1980). Aspects of the bonobo procedure may have contributed to this failure. Most important, no response was required in the presence of the stimuli to encourage attention to them. Furthermore, learning to make one response to the target sequence and another to the other sequences involves a bias that allows for better than chance responding. With the two-alternative forced-choice procedure used with the bonobos, the R1 response is correct for one sequence, whereas the R2 response is correct for three sequences. To correct for this, there are three times as many AB trials as each of the other sequences. However, this correction allows a bias to develop in which reinforcement often can be obtained by using only the last stimulus seen as the basis of choice (e.g., when the last stimulus is B respond R1 when the last stimulus is A respond R2). This solution yields reinforcement on five out of six, or 83%, of the trials. In the present experiment with pigeons, using this two-alternative forced choice procedure, most subjects tended to base their choice on the last-seen stimulus. This design allowed subjects to use a suboptimal but relatively effective choice strategy.

Determining key residues of engineered scFv antibody variants with improved MMP-9 binding using deep sequencing and machine learning

Given the crucial role of specific matrix metalloproteinases (MMPs) in the extracellular matrix, an imbalance in the regulation of activation of matrix metalloproteinase-9 (MMP-9) zymogen and inhibition of the enzyme can result in various diseases, such as cancer, neurodegenerative, and gynecological diseases. Thus, developing novel therapeutics that target MMP-9 with single-chain antibody fragments (scFvs) is a promising approach. We used fluorescent-activated cell sorting (FACS) to screen a synthetic scFv antibody library displayed on yeast for enhanced binding to MMP-9. The screened scFv mutants demonstrated improved binding to MMP-9 compared to the natural inhibitor of MMPs, tissue inhibitor of metalloproteinases (TIMPs). To identify the molecular determinants of these engineered scFv variants that affect binding to MMP-9, we used next-generation DNA sequencing and computational protein structure analysis. Additionally, a deep-learning language model was trained on the screened scFv library of variants to predict the binding affinities of scFv variants based on their CDR-H3 sequences.

Spatiotemporal coupling deep neural network for time-resolved flow field reconstruction around a circular cylinder

We propose a spatiotemporal coupling deep neural network approach for time-resolved reconstruction of the velocity field around a circular cylinder. The neural network leverages two distinct data types: (1) non-time-resolved velocity field around the cylinder, consisting of fixed frequency sampling and variable frequency sampling velocity field, and (2) the time-resolved surface pressure sequence around the cylinder. The deep neural network comprises two sub-networks: a convolutional autoencoder (CAE) for nonlinear mode extraction and a Transformer for sequence-to-sequence learning. We refer to this architecture as CTNet (CAE-Transformer Network). The encoder in the CAE maps non-time-resolved velocity field to a latent vector, enabling the extraction of nonlinear modal coefficients. An appropriate time window length for the surface pressure sequence is then selected to establish a Transformer sequence learning model, using the chosen sequence as input to predict the corresponding nonlinear modal coefficients. Once the Transformer is well trained, the time-resolved nonlinear modal coefficients of velocity field can be achieved. Along with the well-trained decoder in the CAE, the time-resolved velocity field can be reconstructed from the output of the Transformer. We verify the performance of CTNet by a simulated dataset at a representative Reynolds number of 3900. The results show a relative reconstruction error of just 6.3% for the time-resolved velocity field, demonstrating high reliability in the reconstruction. We further compare the reconstructed velocity field obtained with and without the utilization of variable frequency sampling velocity field. Notably, the inclusion of variable frequency sampling velocity field significantly improves the reconstruction quality.

Self-Study and Discussion Promote Students’ Science Learning

AbstractMany have investigated the effectiveness of teaching methods across different disciplines. In a previous study, we found the superior effectiveness of a two-part learning sequence consisting of self-study followed by discussion in humanities education. To examine the possibility of similar findings in science education, we compared that sequence with a traditional lecture–based teaching. Results indicate that self-study combined with discussion yields higher test scores, demonstrating its potential as a beneficial teaching strategy in science education. In Experiment 1, undergraduate students from various disciplines studied introductory biology under five different conditions: self-study and discussion, lecture and discussion, lecture and review, self-study and review, and self-study and review with additional learning materials. Results indicated that the self-study and discussion condition yielded higher test scores than both the lecture and discussion condition and the three review groups. There were no significant differences in scores among the three review groups. These results demonstrate that discussion is an effective learning method and that self-study enhances the benefits of discussion. In Experiment 2, which involved science and engineering majors, we successfully replicated the design of Experiment 1 using physics for the two discussion conditions: self-study or lecture. We also performed in-depth analyses of student interactions during the discussions. Among the two discussion groups, the self-study and discussion group exhibited more active and constructive engagement in sharing unknowns, leading to enhanced learning. These findings suggest that integrating self-study with discussion can significantly enhance learning outcomes in science education. This approach not only improves test performance but also fosters deeper student thinking. Thus, educators might consider adopting this method to create more interactive and effective learning environments in science courses.

Utilizing machine learning and hemagglutinin sequences to identify likely hosts of influenza H3Nx viruses

Influenza is a disease that represents both a public health and agricultural risk with pandemic potential. Among the subtypes of influenza A virus, H3 influenza virus can infect many avian and mammalian species and is therefore a virus of interest to human and veterinary public health. The primary goal of this study was to train and validate classifiers for the identification of the most likely host species using the hemagglutinin gene segment of H3 viruses. A five-step process was implemented, which included training four machine learning classifiers, testing the classifiers on the validation dataset, and further exploration of the best-performing model on three additional datasets. The gradient boosting machine classifier showed the highest host-classification accuracy with a 98.0 % (95 % CI [97.01, 98.73]) correct classification rate on an independent validation dataset. The classifications were further analyzed using the predicted probability score which highlighted sequences of particular interest. These sequences were both correctly and incorrectly classified sequences that showed considerable predicted probability for multiple hosts. This showed the potential of using these classifiers for rapid sequence classification and highlighting sequences of interest. Additionally, the classifiers were tested on a separate swine dataset composed of H3N2 sequences from 1998 to 2003 from the United States of America, and a separate canine dataset composed of canine H3N2 sequences of avian origin. These two datasets were utilized to look at the applications of predicted probability and host convergence over time. Lastly, the classifiers were used on an independent dataset of environmental sequences to explore the host identification of environmental sequences. The results of these classifiers show the potential for machine learning to be used as a host identification technique for viruses of unknown origin on a species-specific level.

Identification of miR-20a as a Diagnostic and Prognostic Biomarker in Colorectal Cancer: MicroRNA Sequencing and Machine Learning Analysis.

The differential expression of miRNAs, a key regulator in many cell signaling pathways, has been studied in various malignancies and may have an important role in cancer progression, including colorectal cancer (CRC). The present study used machine learning and gene interaction study tools to explore the prognostic and diagnostic value of miRNAs in CRC. Integrative analysis of 353 CRC samples and normal tissue data was obtained from the TCGA database and further analyzed by R packages to define the deferentially expressed miRNAs (DEMs). Furthermore, machine learning and Kaplan Meier survival analysis helped better specify the significant prognostic value of miRNAs. A combination of online databases was then used to evaluate the interactions between target genes, their molecular pathways, and the correlation between the DEMs. The results indicated that miR-19b and miR-20a have a significant prognostic role and are associated with CRC progression. The ROC curve analysis discovered that miR-20a alone and combined with other miRNAs, including hsa-mir-21 and hsa-mir-542, are diagnostic biomarkers in CRC. In addition, 12 genes, including NTRK2, CDC42, EGFR, AGO2, PRKCA, HSP90AA1, TLR4, IGF1, ESR1, SMAD2, SMAD4, and NEDD4L, were found to be the highest score targets for these miRNAs. Pathway analysis identified the two correlated tyrosine kinase and MAPK signaling pathways with the key interaction genes, i.e., EGFR, CDC42, and HSP90AA1. To better define the role of these miRNAs, the ceRNA network, including lncRNAs, was also prepared. In conclusion, the combination of R data analysis and machine learning provides a robust approach to resolving complicated interactions between miRNAs and their targets.