Non-negative Matrix Factorization Approach Research Articles

The prognostic value of sialylation-related long non-coding RNAs in lung adenocarcinoma

There has been increasing interest in the role of epigenetic modification in cancers recently. Among the various modifications, sialylation has emerged as a dominant subtype implicated in tumor progression, metastasis, immune evasion, and chemoresistance. The prognostic significance of sialylation-related molecules has been demonstrated in colorectal cancer. However, the potential roles and regulatory mechanisms of sialylation in lung adenocarcinoma (LUAD) have not been thoroughly investigated. Through Pearson correlation, univariate Cox hazards proportional regression, and random survival forest model analyses, we identified several prognostic long non-coding RNAs (lncRNAs) associated with aberrant sialylation and tumor progression, including LINC00857, LINC00968, LINC00663, and ITGA9-AS1. Based on the signatures of four lncRNAs, we classified patients into two clusters with different landscapes using a non-negative matrix factorization approach. Collectively, patients in Cluster 1 (C1) exhibited worse prognoses than those in Cluster 2 (C2), as well as heavier tumor mutation burden. Functional enrichment analysis showed the enrichment of several pro-tumor pathways in C1, differing from the upregulated Longevity and programmed cell death pathways in C2. Moreover, we profiled immune infiltration levels of important immune cell lineages in two subgroups using MCPcounter scores and single sample gene set enrichment analysis scores, revealing a relatively immunosuppressive microenvironment in C1. Risk analysis indicated that LINC00857 may serve as a pro-tumor regulator, while the other three lncRNAs may be protective contributors. Consistently, we observed upregulated LINC00857 in C1, whereas increased expressive levels of LINC00968, LINC00663, and ITGA9-AS1 were observed in C2. Finally, drug sensitivity analysis suggested that patients in the two groups may benefit from different therapeutic strategies, contributing to precise treatment in LUAD. By integrating multi-omics data, we identified four core sialylation-related lncRNAs and successfully established a prognostic model to distinguish patients with different characterizations. These findings may provide some insights into the underlying mechanism of sialylation, and offer a new stratification way as well as clinical guidance in LUAD.

Exploring collaborative filtering through K-Nearest Neighbors and Non-Negative Matrix Factorization

Collaborative filtering (CF) algorithms have received a lot of interest in recommender systems due to their ability to give personalized recommendations by exploiting user-item interaction data. In this article, we explore two popular CF methods—K-Nearest Neighbors (KNN) Regression and Non-Negative Matrix Factorization (NMF)—in detail as we dig into the world of collaborative filtering. Our goal is to evaluate their performance on the MovieLens 1M dataset and offer information about their advantages and disadvantages. A thorough explanation of the significance of recommender systems in contemporary content consumption settings is given at the outset of our examination. We look into Collaborative Filtering's complexities and how it uses user choices to produce tailored recommendations. Then, after setting the scene, we explain the KNN Regression and NMF approaches, going over their guiding principles and how they apply to recommendation systems. We conduct an extensive investigation of KNN Regression and NMF on the MovieLens 1M dataset to provide a thorough evaluation. We describe the model training processes, performance measures, and data pre-processing steps used. We measure and analyse the predicted accuracy of these strategies using empirical studies, revealing light on their effectiveness when applied to various user preferences and content categories.

T helper cells exhibit a dynamic and reversible 3'-UTR landscape.

3' untranslated regions (3' UTRs) are critical elements of messenger RNAs, as they contain binding sites for RNA-binding proteins (RBPs) and microRNAs that affect various aspects of the RNA life cycle including transcript stability and cellular localization. In response to T cell receptor activation, T cells undergo massive expansion during the effector phase of the immune response and dynamically modify their 3' UTRs. Whether this serves to directly regulate the abundance of specific mRNAs or is a secondary effect of proliferation remains unclear. To study 3'-UTR dynamics in T helper cells, we investigated division-dependent alternative polyadenylation (APA). In addition, we generated 3' end UTR sequencing data from naive, activated, memory, and regulatory CD4+ T cells. 3'-UTR length changes were estimated using a nonnegative matrix factorization approach and were compared with those inferred from long-read PacBio sequencing. We found that APA events were transient and reverted after effector phase expansion. Using an orthogonal bulk RNA-seq data set, we did not find evidence of APA association with differential gene expression or transcript usage, indicating that APA has only a marginal effect on transcript abundance. 3'-UTR sequence analysis revealed conserved binding sites for T cell-relevant microRNAs and RBPs in the alternative 3' UTRs. These results indicate that poly(A) site usage could play an important role in the control of cell fate decisions and homeostasis.

Disentangling Multiple Emitting Components in Molecular Observations with Nonnegative Matrix Factorization

Molecular emission from the galactic and extragalactic interstellar medium (ISM) is often used to determine the physical conditions of the dense gas. However, even from spatially resolved regions, the observed molecules do not necessarily arise from a single component. Disentangling multiple gas components is often a degenerate problem in radiative transfer studies. In this paper, we investigate the use of the nonnegative matrix factorization (NMF) approach as a means to recover gas components from a set of blended line intensity maps of molecular transitions that may trace different physical conditions. We run a series of experiments on synthetic data sets designed to replicate conditions in two very different environments: galactic pre-stellar cores and the ISM in high-redshift galaxies. We find that the NMF algorithm often recovers the multiple components resembling those used in the data-generating process, provided that the different components have similar column densities. When NMF fails to recover all the individual components it does however group together the most similarly emitting ones. We further found that initialization and regularisation are key factors in the efficiency of the NMF algorithm.

Application of Weibull Distribution to Hidden Markov Model for Non-Negative Factorization Matrix

Probabilistic Nonnegative Matrix Factorizations (NMFs) are very useful in statistics when dealing with stochastic signals such as wave fronts, share prices, and volatility as a Nonnegative Matrix Factorization (NMF) approach. Little attention has been made in the literature to developing NMF algorithms that use moving average to exploit data's temporal dependencies. A hidden Markov model (HMM) using a Weibull distribution as the output density function was created in this study. The Weibull HMM was then reformulated as a probabilistic NMF. This demonstrates the connection between the proposed HMM and NMF, and will lead to a novel probabilistic NMF approach in which the model captures temporal dependencies inherently utilizing moving average. Furthermore, the model parameters were estimated using maximum likelihood estimation (MLE). The model's adaptability was compared to the existing probabilistic NMFs models of gamma and lognormal. Our trials with US COVID-19 data revealed that the proposed technique achieves a superior balance of sparsity, the goodness of fit, and temporal modeling than gamma and lognormal models.

Establishment of novel ferroptosis-related prognostic subtypes correlating with immune dysfunction in prostate cancer patients

BackgroundWe aimed to identify two new prognostic subtypes and create a predictive index for prostate cancer (PCa) patients based on ferroptosis database. MethodsThe nonnegative matrix factorization approach was used to identify molecular subtypes. We investigate the differences between cluster 1 and cluster 2 in terms of clinical features, functional pathways, tumour stemness, tumour heterogeneity, gene mutation and tumour immune microenvironment score after identifying the two molecular subtypes. Colony formation assay and flow cytometry assay were performed. ResultsThe stratification of two clusters was closely connected to BCR-free survival using the nonnegative matrix factorization method, which was validated in the other three datasets. Furthermore, multivariate Cox regression analysis revealed that this classification was an independent risk factor for patients with PCa. Ribosome, aminoacyl tRNA production, oxidative phosphorylation, and Parkinson's disease-related pathways were shown to be highly enriched in cluster 1. In comparison to cluster 2, patients in cluster 1 exhibited significantly reduced CD4+ T cells, CD8+ T cells, neutrophils, dendritic cells and tumor immune microenvironment scores. Only HHLA2 was more abundant in cluster 1. Moreover, we found that P4HB downregulation could significantly inhibit the colony formation ability and contributed to cell apoptosis of C4-2B and DU145 cell lines. ConclusionsWe discovered two new prognostic subtypes associated with immunological dysfunction in PCa patients based on ferroptosis-related genes and found that P4HB downregulation could significantly inhibit the colony formation ability and contributed to cell apoptosis of PCa cell lines.

Molecular Classification of Relapsed DLBCL Reveals Novel Biologic Subgroups

Background Diffuse large B cell lymphoma (DLBCL) is the most common aggressive lymphoid malignancy in adults. Though standard immunochemotherapy regimens can result in clinical remission and cure in a majority of patients, approximately 30% of patients are primary treatment resistant or eventually relapse (rrDLBCL), and their prognosis is very poor. Recent molecular classification by large scale genomic and transcriptomic profiling of newly diagnosed DLBCL tumors has resulted in a deeper understanding of disease drivers, however, the molecular heterogeneity of rrDLBCL is yet to be fully explored. Therefore, in this study, we used an unsupervised approach to better define transcriptional and genetic programs in rrDLBCL samples hypothesizing that molecular stratification within this high-risk group may lead to a better understanding of this disease, which remains an unmet need, and a more personalized therapeutic approach. Methods Data used in this study included RNA (n=143) and whole exome (n=126) sequencing data from available FFPE tumor samples at the time of a relapse (any line of treatment, r1-r10 relapse timepoints included in analysis, one per patient), consented to the Molecular Epidemiology Resource (n=61), banked in the Mayo Lymphoma Biobank (n=50), or consented to the CC-122-ST-001 clinical trial (n=32, NCT01421524). Unsupervised clustering was performed on protein coding gene expression values using the non-negative matrix factorization (NMF) approach. NMF was carried out by performing 200 runs on cluster sizes from 2 to 6. Cell of origin (COO) was determined using the method described by Reddy et al. The tumor microenvironment was analyzed using CIBERSORTx and Lymphoma Microenvironment Classification (LME). Genetic classification was done using LymphGen. Results In order to identify unique transcriptional programs in rrDLBCL, we implemented NMF and consensus clustering to define an ideal quantity of stable patient clusters. We identified 4 rrDLBCL patient clusters (rrC1-rrC4) as the most stable solution, based on consensus clustering selection criteria of iterative NMF runs. While relapse timepoint was not associated with individual clusters, we did observe significant differences in tumor characteristics between the clusters. COO classification found that rrC1 and rrC4 were enriched for GCB-DLBCL while rrC3 was enriched for ABC-DLBCL classification (P < .001). To explore the biologic programs in each cluster, we used pathFindR, which leverages protein-protein interaction networks to identify disease related pathways, rrC1 and rrC4 were defined by pathways involved in cell cycle and mismatch repair; oxidative phosphorylation in rrC2; and MAPK and NF-kB in rrC3. Next, LME classification was done to define patterns in the tumor microenvironment and we found a significant difference between groups (P < .001), with rrC2 and rrC4 tumors enriched for a depleted TME, while rrC1 and rrC3 had a more immune rich or inflammatory TME. This was supported by CIBERSORTx analysis identifying significant differences (P > .001) in abundance of CD8 T cells, memory CD4 T cells, and M1 macrophages between clusters, with rrC4 having the lowest abundance. We then explored genetic patterns in each cluster. Gene level mutation enrichment analysis showed BCL7A mutation exclusively in C4 (6/6, P = .03). The clusters were not associated with LymphGen classification. Lastly, our recently published high risk signature classification that is associated with early clinical failure and is characterized by metabolic dysregulation, a depleted TME, T53 alterations, and poor outcome, was significantly different between the groups (P = .03) with enrichment for high-risk cases in rrC4. In summary, we show for the first time that rrDLBCL patients can be classified into four gene expression clusters that are associated with distinct pathway, TME, and genetic programs. These clusters should now be tested to learn if they can help select patients for newer therapies for rrDLBCL such as CAR-T and bispecific antibodies.

On the Robustness of Muscle Synergies in the Lower Extremity for Human-Exoskeletons Interaction

Accurately recognizing the user's expected action and converting it into a control signal is critical to the effectiveness of exoskeleton control. Inspired by the natural human movement, an alternative method based on muscle synergies has received increasing attention in myoelectric control in robotic exoskeletons. For the muscle synergies based control system, robust and reliable synergies structures are crucial. Therefore, this paper investigated whether robust muscle synergies could be extracted in the lower limb with three daily activities. Firstly, ten channels of lower extremity surface electromyographic (sEMG) signals were recorded from five healthy subjects during movements. Then, the non-negative matrix factorization (NMF) approach was used to extract the muscle synergies, and the variance accounted for (VAF) method was used to count the number of synergies. Finally, cosine similarity (CS) and synergy stability index (SSI) were used to evaluate the robustness. The resulting high SSI and CS scores demonstrated the robustness and reliability of muscle synergy in typical lower extremity tasks. These findings provide valuable insights for developing muscle synergies based assistive technology for human exoskeletons interaction.

Synchronous Muscle Synergy Evaluation of Jaw Muscle Activities during Chewing at Different Speeds, a Preliminary Study.

Human mastication is a complex and rhythmic biomechanical process regulated by the central nervous system (CNS). Muscle synergies are a group of motor primitives that the CNS may combine to simplify motor control in human movement. This study aimed to apply the non-negative matrix factorization approach to examine the coordination of the masticatory muscles on both sides during chewing. Ten healthy individuals were asked to chew gum at different speeds while their muscle activity was measured using surface electromyography of the right and left masseter and temporalis muscles. Regardless of the chewing speed, two main muscle synergies explained most of the muscle activity variation, accounting for over 98% of the changes in muscle patterns (variance accounted for >98%). The first synergy contained the chewing side masseter muscle information, and the second synergy provided information on bilateral temporalis muscles during the jaw closing. Furthermore, there was robust consistency and high degrees of similarity among the sets of muscle synergy information across different rate conditions and participants. These novel findings in healthy participants supported the hypothesis that all participants in various chewing speed conditions apply the same motor control strategies for chewing. Furthermore, these outcomes can be utilized to design rehabilitation approaches such as biofeedback therapy for mastication disorders.

Mapping the macrostructure and microstructure of the in vivo human hippocampus using diffusion MRI.

The hippocampus is classically divided into mesoscopic subfields which contain varying microstructure that contribute to their unique functional roles. It has been challenging to characterize this microstructure with current magnetic resonance based neuroimaging techniques. In this work, we used diffusion magnetic resonance imaging (dMRI) and a novel surface-based approach in the hippocampus which revealed distinct microstructural distributions of neurite density and dispersion, T1w/T2w ratio as a proxy for myelin content, fractional anisotropy, and mean diffusivity. We used the neurite orientation dispersion and density imaging (NODDI) model optimized for grey matter diffusivity to characterize neurite density and dispersion. We found that neurite dispersion was highest in the cornu ammonis (CA) 1 and subiculum subfields which likely captures the large heterogeneity of tangential and radial fibres, such as the Schaffer collaterals, perforant path, and pyramidal neurons. Neurite density and T1w/T2w were highest in the subiculum and CA3 and lowest in CA1, which may reflect known myeloarchitectonic differences between these subfields. Using a simple logistic regression model, we showed that neurite density, dispersion, and T1w/T2w measures were separable across the subfields, suggesting that they may be sensitive to the known variability in subfield cyto- and myeloarchitecture. We report macrostructural measures of gyrification, thickness, and curvature that were in line with ex vivo descriptions of hippocampal anatomy. We employed a multivariate orthogonal projective non-negative matrix factorization (OPNNMF) approach to capture co-varying regions of macro- and microstructure across the hippocampus. The clusters were highly variable along the medial-lateral (proximal-distal) direction, likely reflecting known differences in morphology, cytoarchitectonic profiles, and connectivity. Finally, we show that by examining the main direction of diffusion relative to canonical hippocampal axes, we could identify regions with stereotyped microstructural orientations that may map onto specific fibre pathways, such as the Schaffer collaterals, perforant path, fimbria, and alveus. These results highlight the value of combining in vivo dMRI with computational approaches for capturing hippocampal microstructure, which may provide useful features for understanding cognition and for diagnosis of disease states.

TSM-CV: Twitter Sentiment Analysis for COVID-19 Vaccines Using Deep Learning

The coronavirus epidemic has imposed a devastating impact on humans around the globe, causing profound anxiety, fear, and complex emotions and feelings. Vaccination against the new coronavirus has started, and people’s feelings are becoming more diverse and complicated. In the presented work, our goal is to use the deep learning (DL) technique to understand and elucidate their feelings. Due to the advancement of IT and internet facilities, people are socially connected to explain their emotions and sentiments. Among all social sites, Twitter is the most used platform among consumers and can assist scientists to comprehend people’s opinions related to anything. The major goal of this work is to understand the audience’s varying sentiments about the vaccination process by using data from Twitter. We have employed both the historic (All COVID-19 Vaccines Tweets Kaggle dataset) and real (tweets) data to analyze the people’s sentiments. Initially, a preprocessing step is applied to the input samples. Then, we use the FastText approach for computing semantically aware features. In the next step, we apply the Valence Aware Dictionary for sentiment Reasoner (VADER) method to assign the labels to the collected features as being positive, negative, or neutral. After this, a feature reduction step using the Non-Negative Matrix Factorization (NMF) approach is utilized to minimize the feature space. Finally, we have used the Random Multimodal Deep Learning (RMDL) classifier for sentiment prediction. We have confirmed through experimentation that our work is effective in examining the emotions of people toward the COVID-19 vaccines. The presented work has acquired an accuracy result of 94.81% which is showing the efficacy of our strategy. Other standard measures like precision, recall, F1-score, AUC, and confusion matrix are also reported to show the significance of our work. The work is aimed to improve public understanding of coronavirus vaccines which can help the health departments to stop the anti-vaccination leagues and motivate people to a booster dose of coronavirus.

Pareto-Optimized Non-Negative Matrix Factorization Approach to the Cleaning of Alaryngeal Speech Signals.

The problem of cleaning impaired speech is crucial for various applications such as speech recognition, telecommunication, and assistive technologies. In this paper, we propose a novel approach that combines Pareto-optimized deep learning with non-negative matrix factorization (NMF) to effectively reduce noise in impaired speech signals while preserving the quality of the desired speech. Our method begins by calculating the spectrogram of a noisy voice clip and extracting frequency statistics. A threshold is then determined based on the desired noise sensitivity, and a noise-to-signal mask is computed. This mask is smoothed to avoid abrupt transitions in noise levels, and the modified spectrogram is obtained by applying the smoothed mask to the signal spectrogram. We then employ a Pareto-optimized NMF to decompose the modified spectrogram into basis functions and corresponding weights, which are used to reconstruct the clean speech spectrogram. The final noise-reduced waveform is obtained by inverting the clean speech spectrogram. Our proposed method achieves a balance between various objectives, such as noise suppression, speech quality preservation, and computational efficiency, by leveraging Pareto optimization in the deep learning model. The experimental results demonstrate the effectiveness of our approach in cleaning alaryngeal speech signals, making it a promising solution for various real-world applications.

Differential Expression of Anomalous Self-Experiences in Spontaneous Speech in Clinical High-Risk and Early-Course Psychosis Quantified by Natural Language Processing

BackgroundBasic self-disturbance, or anomalous self-experiences (ASEs), is a core feature of the schizophrenia spectrum. We propose a novel method of natural language processing to quantify ASEs in spoken language by direct comparison to an inventory of self-disturbance, the Inventory of Psychotic-Like Anomalous Self-Experiences (IPASE). We hypothesized that there would be increased similarity in open-ended speech to the IPASE items in individuals with early-course psychosis (PSY) compared with healthy individuals, with clinical high-risk (CHR) individuals intermediate in similarity. MethodsOpen-ended interviews were obtained from 170 healthy control participants, 167 CHR participants, and 89 PSY participants. We calculated the semantic similarity between IPASE items and “I” sentences from transcribed speech samples using S-BERT (Sentence Bidirectional Encoder Representation from Text). Kolmogorov-Smirnov tests were used to compare distributions across groups. A nonnegative matrix factorization of cosine similarity was performed to rank IPASE items. ResultsSpoken language of CHR individuals had the greatest semantic similarity to IPASE items when compared to both healthy control (s = 0.44, p < 10−14) and PSY (s = 0.36, p < 10−6) individuals, while IPASE scores were higher among PSY than CHR group participants. In addition, the nonnegative matrix factorization approach produced a data-driven domain that differentiated the CHR group from the others. ConclusionsWe found that open-ended interviews elicited language with increased semantic similarity to the IPASE by participants in the CHR group compared with patients with psychosis. This demonstrates the utility of these methods for differentiating patients from healthy control participants. This complementary approach has the capacity to scale to large studies investigating phenomenological features of schizophrenia and potentially other clinical populations.

Non-negative tensor factorization for vibration-based local damage detection

In this study, a novel non-negative tensor factorization (NTF)-based method for vibration-based local damage detection in rolling element bearings is proposed. As the diagnostic signal registered from a faulty machine is non-stationary, the time–frequency method is frequently used as a primary decomposition technique. It is proposed here to extract multi-linear NTF-based components from a 3D array of time–frequency representations of an observed signal partitioned into blocks. As a result, frequency and temporal informative components can be efficiently separated from non-informative ones. The experiments performed on synthetic and real signals demonstrate the high efficiency of the proposed method with respect to the already known non-negative matrix factorization approach.

RRNMF-MAGL: Robust regularization non-negative matrix factorization with multi-constraint adaptive graph learning for dimensionality reduction

In this paper, a new unsupervised dimensionality reduction method named robust regularization non-negative matrix factorization with multi-constraint adaptive graph learning (RRNMF-MAGL) is developed. Compared with the existing methods, RRNMF-MAGL can extract robust low-dimensional features and adaptively learn a manifold structure to well reflect the data distribution. Specifically, to alleviate the influence of outliers and noises, a robust non-negative matrix factorization (RNMF) approach with L21-norm loss function is first constructed. Then, a multi-constraint adaptive graph learning (MAGL) model is designed based on low-dimensional features to reduce the computational complexity and avoid the adverse influence of redundant information. Moreover, multiple constraints (i.e., sparsity and locality) are incorporated into graph learning for enhancing the discrimination ability of graph structure. Next, to make the learned low-dimensional features well maintain the local geometric structure information of data, a graph Laplacian regularization (GLR) is designed. Finally, an iterative update strategy is proposed to optimize RRNMF-MAGL and its convergence is verified by numerical experiments. The superior effectiveness and robustness of RRNMF-MAGL is demonstrated by extensive experiments on eight publicly available datasets.

Non-negative matrix factorization for overlapping community detection in directed weighted networks with sparse constraints.

Detecting overlapping communities is essential for analyzing the structure and function of complex networks. However, most existing approaches only consider network topology and overlook the benefits of attribute information. In this paper, we propose a novel attribute-information non-negative matrix factorization approach that integrates sparse constraints and optimizes an objective function for detecting communities in directed weighted networks. Our algorithm updates the basic non-negative matrix adaptively, incorporating both network topology and attribute information. We also add a sparsity constraint term of graph regularization to maintain the intrinsic geometric structure between nodes. Importantly, we provide strict proof of convergence for the multiplication update rule used in our algorithm. We apply our proposed algorithm to various artificial and real-world networks and show that it is more effective for detecting overlapping communities. Furthermore, our study uncovers the intricate iterative process of system evolution toward convergence and investigates the impact of various variables on network detection. These findings provide insights into building more robust and operable complex systems.

Abstract 299: Large tandem duplications in cancer resulting from transcription and DNA replication collisions

Abstract Somatic structural variations (SVs) are common in cancer. Although a small fraction of SVs in breast and ovarian cancers can be attributed to homologous recombination deficiency, the underlying molecular mechanisms for the vast majority of somatic SVs remain unclear. Here, we focus on the roles of transcription and DNA replication collisions in genomic instability in cancer. Such collisions are unavoidable in cells since both transcription and replication use the same DNA as template. We hypothesized that transcription replication collisions (TRCs), if not properly repaired, would lead to collapsed replication forks and result in SVs. To this end, we studied somatic SVs in 5994 high-coverage whole-genome sequenced primary and metastatic tumors from three independent pan-cancer cohorts. A total of 12 conserved SV signatures, representing independent molecular mechanisms, were deconvoluted from these cohorts using non-negative matrix factorization approach. We detected replicated-strand bias, the expected footprint of transcription-replication collision, in large tandem duplications (TDs) across multiple cohorts. This bias was only observed in expressed genes, consistent with TRCs depending on transcription activity. Large TDs were abundant in female-specific (breast, ovarian and uterus), upper gastric-intestinal tract and prostate cancers. They were associated with worse patient survival and TP53 and CDK12 mutations. CDK12 is a cyclin-dependent kinase (CDK) and a key regulator of transcription elongation. Deleting or suppressing CDK12 using CRISPR-Cas9 in prostate cancer cell lines not only increased RNA:DNA hybrids (R-loops), but also promoted TRCs, suggesting a mechanism by which dysregulation of a transcriptional CDK may lead to genomic instability. Finally, using existing large-scale drug screening data, we found that cancer cell lines with abundant large TDs were significantly more sensitive to the WEE1 inhibitor, MK-1775, which we experimentally validated in prostate cancer cells lacking CDK12. In summary, our data suggest that large TDs in cancer form due to impaired TRC repair and can be used as a biomarker for prognosis and treatment. Citation Format: Yang Yang, Michelle Badura, Patrick O’Leary, Emily Egusa, Troy Robinson, Xiaoming Zhong, Jason Swinderman, Minkyu Kim, Haolong Li, Alan Ashworth, Felix Feng, Jonathan Chou, Lixing Yang. Large tandem duplications in cancer resulting from transcription and DNA replication collisions [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 299.

Abstract 2076: Novel framework for systematically detecting alternative transcript initiation by integrating ATAC-seq and RNA-seq

Abstract Background: Alternative transcription initiation (ATI) has been frequently observed in cancer suggesting that it contributes to the malignant transformation of the cells. However, ATI remains largely unexplored mainly due to the lack of tools for detecting ATI. We propose a computational method for integrating bulk ATAC-seq and bulk RNA-seq to identify ATI and understand their differential usage between tissues. We hypothesize that differential ATAC-seq intensity can be used as a guide for looking for differential promoter usage, which might enable the identification of novel ATI events in a transcript-agnostic way. Methods: We recently published methods for RNA-seq base-level analysis for identifying structural variations in transcripts. Building on this, we developed a supervised sparse non-negative matrix factorization approach that integrates base-level RNA-seq and ATAC-seq, which aims to identify ATIs as well as characterize the latent structure of the underlying isoforms. This method equips many unique features that can be useful for the identification of novel ATIs. Because the method scans an entire collection of DNA accessible regions provided by ATAC-seq, it enables a comprehensive screening of novel ATI candidates that are not limited to the known promoters. Additionally, the uncompressed view of base-level RNA-seq allows us to infer the structure of individual isoforms independently of known gene annotation. The predicted isoforms can be further used to deconvolute base-level RNA-seq of individual cases into the isoforms and thus infer expression levels of each isoform. Results: We applied this method to a sub-cohort (N=350) of TCGA pan-cancer samples in which both ATAC-seq and RNA-seq are available. Empiric comparison to existing methods confirmed known true positive ATIs including important cancer genes such as CDKN2A and ALK. In particular, the method successfully identified the ATIs including some challenging cases such as ATIs located at internal introns or at constitutive exons in other transcripts. The deconvolution analysis applied to the extended cohort to all available ~10,000 TCGA pan-caner samples across 32 tissue types revealed that ATIs were predominantly differentiated across tissue types. By applying the method to a set of transcription regulator genes, we identified ~1% of the genes had ATIs including the novel cancer-specific genes with known ATIs that had not been previously reported as relevant to cancer. Additionally, we provide examples in which the isoform’s function relative to cancer appears mechanistically interesting. Conclusion: We propose a multi-omics integration method that is independent of known gene annotation, enabling a robust identification of ATIs. Our results strongly demonstrate our ability to pick up known as well as novel ATIs that are otherwise difficult to identify by existing methods. Citation Format: Hyo Young Choi, Won-Young Choi, David N. Hayes. Novel framework for systematically detecting alternative transcript initiation by integrating ATAC-seq and RNA-seq [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2076.

Community detection in networks through a deep robust auto-encoder nonnegative matrix factorization

Community detection is a field of research with increasing influence due to its importance in dimensionality reduction and revealing the organizational patterns in networks. A community or cluster in a network reflects the interactions among the objects that are strongly connected and distinguish them from the ones in other clusters. Over the past decades, various algorithms have been developed to detect the community structure in networks. However, the performance of these algorithm is affected by the existence of noise or nonlinear information in the network. Consequently, deep learning algorithms arise to tackle this issue. In this paper, we propose a novel deep robust auto-encoder nonnegative matrix factorization (DRANMF) approach to detect the community structure in networks. In particular, DRANMF consists of a deep structured decoder and encoder components to transform the high-order proximity matrix of the network to and from the cluster membership space, respectively. Moreover, we enhance the robustness of the proposed approach against noise and outliers by adopting the l2,1-norm to formulate the objective function. The experiments on multiple synthetic and real-world networks demonstrate that the proposed DRANMF has a better performance and robustness than other existing methods.

A weighted non-negative matrix factorization approach to predict potential associations between drug and disease

BackgroundAssociations of drugs with diseases provide important information for expediting drug development. Due to the number of known drug-disease associations is still insufficient, and considering that inferring associations between them through traditional in vitro experiments is time-consuming and costly. Therefore, more accurate and reliable computational methods urgent need to be developed to predict potential associations of drugs with diseases.MethodsIn this study, we present the model called weighted graph regularized collaborative non-negative matrix factorization for drug-disease association prediction (WNMFDDA). More specifically, we first calculated the drug similarity and disease similarity based on the chemical structures of drugs and medical description information of diseases, respectively. Then, to extend the model to work for new drugs and diseases, weighted K nearest neighbor was used as a preprocessing step to reconstruct the interaction score profiles of drugs with diseases. Finally, a graph regularized non-negative matrix factorization model was used to identify potential associations between drug and disease.ResultsDuring the cross-validation process, WNMFDDA achieved the AUC values of 0.939 and 0.952 on Fdataset and Cdataset under ten-fold cross validation, respectively, which outperforms other competing prediction methods. Moreover, case studies for several drugs and diseases were carried out to further verify the predictive performance of WNMFDDA. As a result, 13(Doxorubicin), 13(Amiodarone), 12(Obesity) and 12(Asthma) of the top 15 corresponding candidate diseases or drugs were confirmed by existing databases.ConclusionsThe experimental results adequately demonstrated that WNMFDDA is a very effective method for drug-disease association prediction. We believe that WNMFDDA is helpful for relevant biomedical researchers in follow-up studies.