Integrative Analysis of Multimodal Omics Data

Pharmacogenomics is the study of how genes affect a person's response to drugs. Thus, understanding the effect of drug at the molecular level can be helpful in both drug discovery and personalized medicine. Over the years, transcriptome data upon drug treatment has been collected and several databases compiled before drug treatment cancer cell multi-omics data with drug sensitivity (IC50, AUC) or time-series transcriptomic data after drug treatment. However, analyzing transcriptome data upon drug treatment is challenging since more than 20,000 genes interact in complex ways. In addition, due to the difficulty of both time-series analysis and multi-omics integration, current methods can hardly perform analysis of databases with different data characteristics. One effective way is to interpret transcriptome data in terms of well-characterized biological pathways. Another way is to leverage state-of-the-art methods for multi-omics data integration. In this paper, we developed Drug Response analysis Integrating Multi-omics and time-series data (DRIM), an integrative multi-omics and time-series data analysis framework that identifies perturbed sub-pathways and regulation mechanisms upon drug treatment. The system takes drug name and cell line identification numbers or user's drug control/treat time-series gene expression data as input. Then, analysis of multi-omics data upon drug treatment is performed in two perspectives. For the multi-omics perspective analysis, IC50-related multi-omics potential mediator genes are determined by embedding multi-omics data to gene-centric vector space using a tensor decomposition method and an autoencoder deep learning model. Then, perturbed pathway analysis of potential mediator genes is performed. For the time-series perspective analysis, time-varying perturbed sub-pathways upon drug treatment are constructed. Additionally, a network involving transcription factors (TFs), multi-omics potential mediator genes, and perturbed sub-pathways is constructed, and paths to perturbed pathways from TFs are determined by an influence maximization method. To demonstrate the utility of our system, we provide analysis results of sub-pathway regulatory mechanisms in breast cancer cell lines of different drug sensitivity. DRIM is available at: http://biohealth.snu.ac.kr/software/DRIM/.

Computational integrative analysis has become a significant approach in the data-driven exploration of biological problems. Many integration methods for cancer subtyping have been proposed, but evaluating these methods has become a complicated problem due to the lack of gold standards. Moreover, questions of practical importance remain to be addressed regarding the impact of selecting appropriate data types and combinations on the performance of integrative studies. Here, we constructed three classes of benchmarking datasets of nine cancers in TCGA by considering all the eleven combinations of four multi-omics data types. Using these datasets, we conducted a comprehensive evaluation of ten representative integration methods for cancer subtyping in terms of accuracy measured by combining both clustering accuracy and clinical significance, robustness, and computational efficiency. We subsequently investigated the influence of different omics data on cancer subtyping and the effectiveness of their combinations. Refuting the widely held intuition that incorporating more types of omics data always produces better results, our analyses showed that there are situations where integrating more omics data negatively impacts the performance of integration methods. Our analyses also suggested several effective combinations for most cancers under our studies, which may be of particular interest to researchers in omics data analysis.

Finding disease genes related to cancer is of great importance for diagnosis and treatment. With the development of high-throughput technologies, more and more multiple-level omics data have become available. Thus, it is urgent to develop computational methods to identify cancer genes by integrating these data. We propose an integrative rank-based method called iRank to prioritize cancer genes by integrating multi-omics data in a unified network-based framework. The method was used to identify the disease genes of hepatocellular carcinoma (HCC) in humans using the multi-omics data for HCC from TCGA after building up integrated networks in the corresponding molecular levels. The kernel of iRank is based on an improved PageRank algorithm with constraints. To demonstrate the validity and the effectiveness of the method, we performed experiments for comparison between single-level omics data and multiple omics data as well as with other algorithms: random walk (RW), random walk with restart on heterogeneous network (RWH), PRINCE and PhenoRank. We also performed a case study on another cancer, prostate adenocarcinoma (PRAD). The results indicate the effectiveness and efficiency of iRank which demonstrates the significance of integrating multi-omics data and multiplex networks in cancer gene prioritization.

Cancer heterogeneity presents a major obstacle to effective drug treatment, emphasizing the need for personalized approaches that can accurately predict drug responses. Advances in high-throughput technologies have driven precision medicine initiatives toward integrating multi-omics data, enabling a more comprehensive understanding of tumor biology. However, integration of diverse omics layers poses challenges for computational modeling, as many traditional machine learning (ML) and statistical methods are not designed to capture complex, high-dimensional and multimodal data. This review examines the studies that integrate multi-omics datasets, aiming to enhance drug response prediction (DRP). Specifically, it outlines the most used omics types and computational approaches – classical ML models, as well as advanced deep learning and multimodal integration frameworks for improving DRP, detailing key methodologies and evaluation metrics, such as area under the dose–response curve, F1 score and mean square error, which assess model performance. By summarizing the integrated omics data, computational methods and challenges encountered, this review provides an in-depth overview of the existing landscape of precision medicine and future directions for advancing drug-response prediction.

A survey on single and multi omics data mining methods in cancer data classification.

Ovarian aging occurs earlier than the aging of many other organs and has a lasting impact on women's overall health and well-being. However, effective interventions to slow ovarian aging remain limited, primarily due to an incomplete understanding of the underlying molecular mechanisms and drug targets. Recent advances in omics data resources, combined with innovative computational tools, are offering deeper insight into the molecular complexities of ovarian aging, paving the way for new opportunities in drug discovery and development. This review aims to synthesize the expanding multi-omics data, spanning genome, transcriptome, proteome, metabolome, and microbiome, related to ovarian aging, from both tissue-level and single-cell perspectives. We will specially explore how the analysis of these emerging omics datasets can be leveraged to identify novel drug targets and guide therapeutic strategies for slowing and reversing ovarian aging. We conducted a comprehensive literature search in the PubMed database using a range of relevant keywords: ovarian aging, age at natural menopause, premature ovarian insufficiency (POI), diminished ovarian reserve (DOR), genomics, transcriptomics, epigenomics, DNA methylation, RNA modification, histone modification, proteomics, metabolomics, lipidomics, microbiome, single-cell, genome-wide association studies (GWAS), whole-exome sequencing, phenome-wide association studies (PheWAS), Mendelian randomization (MR), epigenetic target, drug target, machine learning, artificial intelligence (AI), deep learning, and multi-omics. The search was restricted to English-language articles published up to September 2024. Multi-omics studies have uncovered key mechanisms driving ovarian aging, including DNA damage and repair deficiencies, inflammatory and immune responses, mitochondrial dysfunction, and cell death. By integrating multi-omics data, researchers can identify critical regulatory factors and mechanisms across various biological levels, leading to the discovery of potential drug targets. Notable examples include genetic targets such as BRCA2 and TERT, epigenetic targets like Tet and FTO, metabolic targets such as sirtuins and CD38+, protein targets like BIN2 and PDGF-BB, and transcription factors such as FOXP1. The advent of cutting-edge omics technologies, especially single-cell technologies and spatial transcriptomics, has provided valuable insights for guiding treatment decisions and has become a powerful tool in drug discovery aimed at mitigating or reversing ovarian aging. As technology advances, the integration of single-cell multi-omics data with AI models holds the potential to more accurately predict candidate drug targets. This convergence offers promising new avenues for personalized medicine and precision therapies, paving the way for tailored interventions in ovarian aging. Not applicable.

BackgroundHigh-throughput (omic) data have become more widespread in both quantity and frequency of use, thanks to technological advances, lower costs and higher precision. Consequently, computational scientists are confronted by two parallel challenges: on one side, the design of efficient methods to interpret each of these data in their own right (gene expression signatures, protein markers, etc.) and, on the other side, realization of a novel, pressing request from the biological field to design methodologies that allow for these data to be interpreted as a whole, i.e. not only as the union of relevant molecules in each of these layers, but as a complex molecular signature containing proteins, mRNAs and miRNAs, all of which must be directly associated in the results of analyses that are able to capture inter-layers connections and complexity.ResultsWe address the latter of these two challenges by testing an integrated approach on a known cancer benchmark: the NCI-60 cell panel. Here, high-throughput screens for mRNA, miRNA and proteins are jointly analyzed using factor analysis, combined with linear discriminant analysis, to identify the molecular characteristics of cancer. Comparisons with separate (non-joint) analyses show that the proposed integrated approach can uncover deeper and more precise biological information. In particular, the integrated approach gives a more complete picture of the set of miRNAs identified and the Wnt pathway, which represents an important surrogate marker of melanoma progression. We further test the approach on a more challenging patient-dataset, for which we are able to identify clinically relevant markers.ConclusionsThe integration of multiple layers of omics can bring more information than analysis of single layers alone. Using and expanding the proposed integrated framework to integrate omic data from other molecular levels will allow researchers to uncover further systemic information. The application of this approach to a clinically challenging dataset shows its promising potential.

Cancer is one of the leading causes of mortality worldwide, and its inherent diversity and heterogeneity pose significant challenges in early diagnosis, drug development, and prognosis. Accurate identification of cancer subtypes has therefore become a critical aspect of personalized cancer treatment. With the advancement of omics technologies, multi-omics data offer a more comprehensive understanding of cancer's underlying mechanisms. However, effectively integrating such diverse datasets to identify cancer subtypes remains a significant challenge. In this study, we introduce a novel approach, termed HRAGNN, for cancer subtype identification through the integration of multi-omics data. HRAGNN first constructs an integrated graph and then leverages Relational Attention Mechanism (RAM) and Graph Neural Network (GNN) to capture complex features across these multi-omics layers. Subsequently, the Multi-view Fusion Network (MVFN) is employed to fuse the features derived from the different omics data. We evaluated the performance of HRAGNN on three datasets, comparing it with other existing methods. The experimental results demonstrate that HRAGNN outperforms other approaches in terms of several key evaluation metrics.

Breast cancer (BC) is the second leading cause of cancer-related death among women in the United States. Previous studies have shown that African American (AA) women have disproportionately high mortality rate in BC compared to European American (EA) women in the United States. Existing studies have suggested that transfer learning could reduce health disparities in BC by transferring information from a majority group (e.g., EA women) to a minority group (e.g., AA women). However, their performance may be decreased due to limited training patient samples and using single-omics data only. We recently presented a transfer learning method by integrating two multi-omics data for reducing cancer disparities. However, the integration model was not optimized, and its performance in reducing disparities was not robust. To address these concerns, we propose a novel weighted multi-modal transfer learning model that can optimize the integration of multi-omics data to alleviate racial disparities in BC. Specifically, we first calculated the patient-patient similarity by using the Pearson Correlation Coefficient (PCC), based on which we performed weighted integration of multi-omics data. Then, we performed a nested grid search method to find the optimal weights for each type of omics data, which were subsequently fed into a transfer learning model. To demonstrate the effectiveness of our model, we performed the tasks of predicting Disease-Specific Survival (DSS) prognosis for the 1098 female BC patient samples from The Cancer Genome Atlas (TCGA) database. Results suggested that our model with optimized weighted integration of three multi-omics (including mRNA, miRNA and methylation) data significantly outperformed previous multi-omics transfer learning models in terms of AUC. In addition, our multi-modal transfer learning model performed remarkably better than single-omics and two-omics-integration transfer learning models, suggesting that the inclusion of more diverse omics data contributes to a better reduction of racial disparities in BC. Furthermore, the performance of our transfer learning model was also superior to that of conventional mixed models and independent models. In summary, our proposed weighted multi-modal transfer learning model achieves far better performance than existing state-of-the-art methods for reducing health disparities in BC. We expect that our proposed model will provide a new way to reduce health disparities in BC diagnosis, prognosis, and treatment, and it can be extensible to alleviate health disparities in other types of cancer. Citation Format: Min-jeong Baek, Lusheng Li, Jieqiong Wang, Vimla Band, Shibiao Wan. A weighted multi-modal transfer learning model for alleviating racial disparities in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3620.

Computational tools for multiomics data integration have usually been designed for unsupervised detection of multiomics features explaining large phenotypic variations. To achieve this, some approaches extract latent signals in heterogeneous data sets from a joint statistical error model, while others use biological networks to propagate differential expression signals and find consensus signatures. However, few approaches directly consider molecular interaction as a data feature, the essential linker between different omics data sets. The increasing availability of genome-scale interactome data connecting different molecular levels motivates a new class of methods to extract interactive signals from multiomics data. Here we developed iOmicsPASS, a tool to search for predictive subnetworks consisting of molecular interactions within and between related omics data types in a supervised analysis setting. Based on user-provided network data and relevant omics data sets, iOmicsPASS computes a score for each molecular interaction, and applies a modified nearest shrunken centroid algorithm to the scores to select densely connected subnetworks that can accurately predict each phenotypic group. iOmicsPASS detects a sparse set of predictive molecular interactions without loss of prediction accuracy compared to alternative methods, and the selected network signature immediately provides mechanistic interpretation of the multiomics profile representing each sample group. Extensive simulation studies demonstrate clear benefit of interaction-level modeling. iOmicsPASS analysis of TCGA/CPTAC breast cancer data also highlights new transcriptional regulatory network underlying the basal-like subtype as positive protein markers, a result not seen through analysis of individual omics data.

e21016 Background: Advances in high throughput measurement technologies (-omics data) have made it possible to generate high complexity, high volume data for oncology research. Researchers are often confronted many more measurements than samples (p > > > n), which poses challenges for both modeling the complexity of disease at the molecular mechanism level, and overfitting when generating predictive models with complex data. Here, we applied a prior knowledge-driven approach to characterize and classify heavy versus light smokers with lung cancer from The Cancer Genome Atlas, an open source repository that catalogs, harmonizes and hosts -omics data collected from samples generously donated from cancer patients. Methods: We applied a reverse inferencing approach to systematically interrogate RNAseq measurements from tumor and control biopsies against a knowledgebase of directed gene networks curated from published experiments. If patterns observed in the data are significantly similar to those in a network, an inference about the directional activity of that network can be made; e.g., the increased transcriptional activity of NFKB. Our library was nucleated through an open sourced knowledge graph and enhanced with updated and relevant knowledge using the open sourced Biological Expression Language framework. Directed networks were either qualitatively scored and used to build disease models, or semi-quantitatively scored and used as classification features. Results: In LUAD tumors, we detected a pattern of gene signatures which indicated a tumor stem cell-like phenotype characterized by predicted decreases in the activity of pro-differentiation factors and an increased response to hypoxia. Analysis of patients with heavy ( > 40) versus light ( < 10) pack-year burden suggested an augmented dedifferentiation profile in heavy smokers. In this example, improved classification was observed through features compression through directed network scoring compared to using individual RNA measurements selected by filtration methods. Conclusions: In-silico analysis of lung cancer patient biopsies generated hypotheses implicating stem cell signaling in tumors, and a further stratification of this signal based on patient pack year burden. Mechanistic modeling may be a useful application to the overfitting problem often encountered with -omics data in translational studies. Data from other TCGA indications can be used to evaluate the consistency of this type of approach

BackgroundOne major goal of large-scale cancer omics study is to identify molecular subtypes for more accurate cancer diagnoses and treatments. To deal with high-dimensional cancer multi-omics data, a promising strategy is to find an effective low-dimensional subspace of the original data and then cluster cancer samples in the reduced subspace. However, due to data-type diversity and big data volume, few methods can integrative and efficiently find the principal low-dimensional manifold of the high-dimensional cancer multi-omics data.ResultsIn this study, we proposed a novel low-rank approximation based integrative probabilistic model to fast find the shared principal subspace across multiple data types: the convexity of the low-rank regularized likelihood function of the probabilistic model ensures efficient and stable model fitting. Candidate molecular subtypes can be identified by unsupervised clustering hundreds of cancer samples in the reduced low-dimensional subspace. On testing datasets, our method LRAcluster (low-rank approximation based multi-omics data clustering) runs much faster with better clustering performances than the existing method. Then, we applied LRAcluster on large-scale cancer multi-omics data from TCGA. The pan-cancer analysis results show that the cancers of different tissue origins are generally grouped as independent clusters, except squamous-like carcinomas. While the single cancer type analysis suggests that the omics data have different subtyping abilities for different cancer types.ConclusionsLRAcluster is a very useful method for fast dimension reduction and unsupervised clustering of large-scale multi-omics data. LRAcluster is implemented in R and freely available via http://bioinfo.au.tsinghua.edu.cn/software/lracluster/.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-015-2223-8) contains supplementary material, which is available to authorized users.

Integrating multi-omics data from diverse platforms to accurately identify cancer subtypes with similar molecular profiles remains a key challenge in precision medicine. To address this, we introduce a novel approach, Multi-omics Data Low-Rank Representation and Data Merging (MDLRDM), designed to capture both the local and global structures within multi-omics datasets. This method aims to enhance the accuracy and robustness of cancer subtype classification. MDLRDM consists of a two-stage process. In the first stage, a hierarchical weighted matrix is constructed to group samples of the same subtype within each omics data. A low-rank representation is then integrated into this matrix to generate a similarity matrix that reflects the intrinsic relationships between samples. In the second stage, these individual similarity matrices are merged to create a consensus matrix, which is then subjected to spectral clustering for the final identification of cancer subtypes. We evaluated the performance of MDLRDM by comparing it with six state-of-the-art multi-omics integration methods using eight real cancer datasets. The results show that MDLRDM outperforms existing methods in terms of both accuracy and generalizability, providing a more reliable solution for cancer subtype identification. Moreover, the identified subtypes reveal crucial molecular differences that improve our understanding of cancer heterogeneity. These insights not only advance our knowledge of cancer biology but also uncover potential therapeutic targets for clinical intervention.

For many complex heterogeneous diseases, the underlying mechanisms contributing to pathogenesis is still unknown. Although prediction models for complex heterogeneous diseases have been trained on clinical features and imaging data, they may not reveal mechanisms contributing to pathogenesis. The development of human diseases often involves a highly interactive and dynamic system of molecular layers. Networks are a useful framework for representing cross-talk relationships of molecules in biological and disease pathways. Understanding complex biological networks is a fundamental and challenging issue in computational and systems biology. Constructing disease-specific networks is still lacking in many diseases even with current advances. It necessitates the development of methods to integrate protein and genetic interaction information from knowledgebases with omics data to reconstruct networks and develop predictive models to effectively integrate different omics data and network information while being interpretable. Biological networks are often inferred through Gaussian Graphical Models (GGMs) using gene or protein expression data only. GGMs identify conditional dependence by estimating a precision matrix between genes or proteins. However, conventional GGM approaches often ignore prior knowledge about protein-protein interactions (PPI). Recently, several groups have extended GGM to incorporate PPI information. However, these methods are either computationally intractable for large-scale data, or disregard weights in the PPI networks. To address these shortcomings, we developed an augmented high-dimensional graphical lasso (AhGlasso) method to incorporate edge weights in known protein-protein interactions with omics data for global network learning. This new method outperforms weighted graphical lasso-based algorithms with respect to computational time in simulated large-scale data settings while achieving better or comparable prediction accuracy of node connections. Using proteomic data from a study on chronic obstructive pulmonary disease (COPD), we demonstrated that AhGlasso improves protein network inference compared to other approaches that incorporate PPI information. Recently Deep Learning including convolution Graph Neural Network (ConvGNN) has shown great potential for disease classification by integrating a known PPI network from existing databases and a single omics data set. We extended the existing ConvGNN framework to incorporate PPI in addition to multi-omics data to develop a prediction model for COPD. This model achieved significantly better accuracy compared to several conventional classification methods. We also demonstrated that the updated COPD-associated network with AhGlasso further improves the prediction accuracy. Moreover, we applied explainable machine learning methods for interpretation and identified significant genes/proteins and important subnetworks in the ConvGNN model for COPD prediction. Our study suggests the benefits of using ConvGNN for the prediction of heterogeneous complex diseases by integrating multi-omics data and interaction networks.

Cell clustering is foundational for analyzing the heterogeneity of biological tissues using single-cell sequencing data. With the maturation of single-cell multi-omics sequencing technologies, we can integrate multiple omics data to perform cell clustering, thereby overcoming the limitations of insufficient information from single omics data. Existing methods for cell clustering often only consider the differences in data patterns during the analysis of multi-omics data, but the dependencies between omics features of different cell types also significantly influence cell clustering. Moreover, the high dropout rates in scRNA-seq and scATAC-seq data can impact the performance of cell clustering. We propose a cell clustering model based on a masked autoencoder, scDRMAE. Utilizing a masking mechanism, scDRMAE effectively learns the relationships between different features and imputes false zeros caused by dropout events. To differentiate the importance of various omics data in cell clustering, we dynamically adjust the weights of different omics data through an attention mechanism. Finally, we use the K-means algorithm for cluster analysis of the fused multi-omics data. On commonly used sets of 15 multi-omics datasets, our method demonstrates superior cell clustering performance on multiple metrics compared to other computational methods. In addition, when datasets exhibit varying degrees of dropout noise, our method shows better performance and stronger stability on multiple metrics compared to other methods. Moreover, by analyzing the cell clusters classified by scDRMAE, we identified several biologically significant biomarkers that have been validated, further confirming the effectiveness of scDRMAE in cell clustering from a biological perspective.