Analysis of cancer omics data: a selective review of statistical techniques

Combining multi-omics analysis with machine learning to uncover novel molecular subtypes, prognostic markers, and insights into immunotherapy for melanoma

BACKGROUND

Melanoma (SKCM) is an extremely aggressive form of cancer, characterized by high mortality rates, frequent metastasis, and limited treatment options. Our study aims to identify key target genes and enhance the diagnostic accuracy of melanoma prognosis by employing multi-omics analysis and machine learning techniques, ultimately leading to the development of novel therapeutic strategies.

METHODS

We obtained and processed transcriptomic data, including RNA expression profiles, methylation microarray data, gene mutation data, and clinical information, from the TCGA dataset using multi-omics analysis and machine learning techniques. We comprehensively evaluated the molecular subtypes of melanoma, the characteristics of the tumor microenvironment (TME), and their effects on patient outcomes. By analyzing the TCGA-SKCM and GEO cohorts, we identified three melanoma subtypes with distinct prognostic features. Additionally, we developed a machine learning-driven signature (MLDS) based on marker genes for different molecular subtypes to significantly improve the prognostic prediction accuracy for melanoma patients. We also extensively examined differences in clinical features, immune cell infiltration, mutational landscapes, and drug treatment effects between high- and low-scoring subgroups. The predictive reliability of MLDS was further explored by knocking down the key signature gene AGPAT2 in melanoma cells using small interfering RNA.

RESULTS

The MLDS demonstrated high C-index values in both the training and validation cohorts, indicating its potential for clinical decision-making. The study also found that MLDS scores were associated with reduced immune cell infiltration and lower expression levels of immune checkpoints. Patients in the low MLDS group may be more responsive to chemotherapeutic agents and more likely to benefit from immune checkpoint inhibitors (ICIs). Single-cell sequencing analysis revealed that the MAPK signaling pathway between AGPAT2 + melanoma cells and fibroblasts/myeloid cells promotes tumor survival in the TME. Finally, the oncogenic role of AGPAT2 in melanoma cell lines was successfully confirmed through cell function assays and subcutaneous tumor formation assays in nude mice.

CONCLUSION

This study not only uncovers the diversity and complexity of melanoma molecular subtypes but also underscores the crucial role of the TME in melanoma progression. It provides new insights and tools for personalized treatment and prognostic assessment of SKCM.

A novel tool for multi-omics network integration and visualization: A study of glioma heterogeneity

Gliomas are highly heterogeneous tumors with generally poor prognoses. Leveraging multi-omics data and network analysis holds great promise in uncovering crucial signatures and molecular relationships that elucidate glioma heterogeneity. However, the complexity of the problem and the high dimensionality of the data increase the challenges of integrating information across various biological levels. This study develops a comprehensive framework aimed at identifying potential glioma-type-specific biomarkers through innovative variable selection and integrated network visualization. We designed a two-step framework for variable selection using sparse network estimation across various omics datasets. This framework incorporates MINGLE (Multi-omics Integrated Network for GraphicaL Exploration), a novel methodology designed to merge distinct multi-omics information into a single network, enabling the identification of underlying relations through an innovative integrated visualization. The analysis was conducted using glioma omics datasets, with patients grouped based on the latest glioma classification guidelines. Our investigation of the glioma data led to the identification of variables potentially serving as glioma-type-specific biomarkers. The integration of multi-omics data into a single network through MINGLE facilitated the discovery of molecular relationships that reflect glioma heterogeneity, supporting the biological interpretation. Scripts and files for reproducing the analysis or adapting it to other applications, are available in R software.

Identification of prognostic subtypes and the role of FXYD6 in ovarian cancer through multi-omics clustering

Background

Ovarian cancer (OC), as a malignant tumor that seriously endangers the lives and health of women, is renowned for its complex tumor heterogeneity. Multi-omics analysis, as an effective method for distinguishing tumor heterogeneity, can more accurately differentiate the prognostic subtypes with differences among patients with OC. The aim of this study is to explore the prognostic subtypes of OC and analyze the molecular characteristics among the different subtypes.

Methods

We utilized 10 clustering algorithms to analyze the multi-omics data of OC patients from The Cancer Genome Atlas (TCGA). After that, we integrated them with ten different machine-learning methods in order to determine high-resolution molecular subgroups and generate machine-learning-driven characteristics that are both resilient and consensus-based. Following the application of multi-omics clustering, we were able to identify two cancer subtypes (CSs) that were associated with the prognosis. Among these, CS2 demonstrated the most positive predictive outcome. Subsequently, five genes that constitute the machine learning (ML)-driven features were screened out by ML algorithms, and these genes possess a powerful predictive ability for prognosis. Subsequently, the function of FXYD Domain-Containing Ion Transport Regulator 6 (FXYD6) in OC was analyzed through gene knockdown and overexpression, and the mechanism by which it affects the functions of OC was explored.

Results

Through multi-omics analysis, we ascertained that the high-risk score group exhibits a poorer prognosis and lack of response to immunotherapy. Moreover, this group is more prone to display the "cold tumor" phenotype, with a lower likelihood of benefiting from immunotherapy. FXYD6, being a crucial differential molecule between subtypes, exerts a tumor-promoting effect when knocked down; conversely, its overexpression yields an opposite outcome. Additionally, we discovered that the overexpression of FXYD6 can induce ferroptosis in OC cells, implying that a low level of FXYD6 in OC cells can safeguard them from ferroptosis. Insightful and more precise molecular categorization of OC can be achieved with a thorough examination of multi-omics data. There are significant consequences for clinical practice stemming from the discovery of risk scores since they provide a useful tool for early prognosis prediction as well as the screening of candidates for immunotherapy.

Machine learning and multi-omics characterization of SLC2A1 as a prognostic factor in hepatocellular carcinoma: SLC2A1 is a prognostic factor in HCC

Hepatocellular carcinoma (HCC) is characterized by high incidence, significant mortality, and marked heterogeneity, making accurate molecular subtyping essential for effective treatment. Using multi-omics data from HCC patients, we applied diverse clustering algorithms to identify three HCC subtypes (HSs) with distinct prognostic characteristics. Among these, HS1 emerged as an immune-compromised subtype associated with the poorest prognosis. Additionally, we developed a novel, robust, and highly accurate machine learning-guided prognostic signature (MLPS) by integrating multiple machine learning algorithms and their combinations. Our study also identified SLC2A1, the core gene of MLPS, as being highly expressed during advanced stages of tumor progression. Knockdown experiments demonstrated that reducing SLC2A1 expression significantly suppressed the malignant behavior of HCC cells. Furthermore, SLC2A1 expression was linked to responsiveness to dasatinib and vincristine, suggesting potential therapeutic relevance. MLPS and SLC2A1 offer promising tools for individualized prognosis prediction and targeted therapy in HCC, providing new opportunities to improve patient outcomes.

Integrated multi-omics analysis identifies a machine learning-derived signature for predicting prognosis and therapeutic vulnerability in clear cell renal cell carcinoma

AIMS

Clear cell renal cell carcinoma (ccRCC) shows considerable variation within and between tumors, presents varying treatment responses among patients, possibly due to molecular distinctions. This study utilized a multi-center and multi-omics analysis to establish and validate a prognosis and treatment vulnerability signature (PTVS) capable of effectively predicting patient prognosis and drug responsiveness.

MATERIALS AND METHODS

To address this complexity, we constructed an integrative multi-omics analysis using 10 clustering algorithms on ccRCC patient data. Afterwards, we applied bootstrapping in univariate Cox regression and the Boruta algorithm to pinpoint clinically relevant genes. Based on this, we developed a robust PTVS using seven machine learning algorithms.

KEY FINDINGS

Our analysis revealed two distinct ccRCC subtypes with differential prognostic implications, notably identifying subtype 2 with poorer outcomes. Patients in the low PTVS group exhibited superior prognosis statistics and an augmented sensitivity to immunotherapy, features consistent with a 'hot tumor' phenotype. Conversely, individuals within the high PTVS group exhibited diminished prognosis statistic and restricted advantages from immunotherapy. Importantly, the PTVS holds future potential as a notable biomarker for guiding personalized treatment strategies, with four prospective targets (CTSK, XDH, PKMYT1, and EGLN2) indicating therapeutic promise in patients scoring high on PTVS.

SIGNIFICANCE

The integrative analysis of multi-omics data profoundly enhances the molecular stratification of ccRCC, underscoring far-reaching impact of such comprehensive profiling on its therapeutic strategies.

Integrative multi-omic and machine learning approach for prognostic stratification and therapeutic targeting in lung squamous cell carcinoma

The proliferation, metastasis, and drug resistance of cancer cells pose significant challenges to the treatment of lung squamous cell carcinoma (LUSC). However, there is a lack of optimal predictive models that can accurately forecast patient prognosis and guide the selection of targeted therapies. The extensive multi-omic data obtained from multi-level molecular biology provides a unique perspective for understanding the underlying biological characteristics of cancer, offering potential prognostic indicators and drug sensitivity biomarkers for LUSC patients. We integrated diverse datasets encompassing gene expression, DNA methylation, genomic mutations, and clinical data from LUSC patients to achieve consensus clustering using a suite of 10 multi-omics integration algorithms. Subsequently, we employed 10 commonly used machine learning algorithms, combining them into 101 unique configurations to design an optimal performing model. We then explored the characteristics of high- and low-risk LUSC patient groups in terms of the tumor microenvironment and response to immunotherapy, ultimately validating the functional roles of the model genes through in vitro experiments. Through the application of 10 clustering algorithms, we identified two prognostically relevant subtypes, with CS1 exhibiting a more favorable prognosis. We then constructed a subtype-specific machine learning model, LUSC multi-omics signature (LMS) based on seven key hub genes. Compared to previously published LUSC biomarkers, our LMS score demonstrated superior predictive performance. Patients with lower LMS scores had higher overall survival rates and better responses to immunotherapy. Notably, the high LMS group was more inclined toward "cold" tumors, characterized by immune suppression and exclusion, but drugs like dasatinib may represent promising therapeutic options for these patients. Notably, we also validated the model gene SERPINB13 through cell experiments, confirming its role as a potential oncogene influencing the progression of LUSC and as a promising therapeutic target. Our research provides new insights into refining the molecular classification of LUSC and further optimizing immunotherapy strategies.

Review: Cancer and neurodevelopmental disorders: multi-scale reasoning and computational guide

The connection and causality between cancer and neurodevelopmental disorders have been puzzling. How can the same cellular pathways, proteins, and mutations lead to pathologies with vastly different clinical presentations? And why do individuals with neurodevelopmental disorders, such as autism and schizophrenia, face higher chances of cancer emerging throughout their lifetime? Our broad review emphasizes the multi-scale aspect of this type of reasoning. As these examples demonstrate, rather than focusing on a specific organ system or disease, we aim at the new understanding that can be gained. Within this framework, our review calls attention to computational strategies which can be powerful in discovering connections, causalities, predicting clinical outcomes, and are vital for drug discovery. Thus, rather than centering on the clinical features, we draw on the rapidly increasing data on the molecular level, including mutations, isoforms, three-dimensional structures, and expression levels of the respective disease-associated genes. Their integrated analysis, together with chromatin states, can delineate how, despite being connected, neurodevelopmental disorders and cancer differ, and how the same mutations can lead to different clinical symptoms. Here, we seek to uncover the emerging connection between cancer, including pediatric tumors, and neurodevelopmental disorders, and the tantalizing questions that this connection raises.

Advancing lung adenocarcinoma prognosis and immunotherapy prediction with a multi-omics consensus machine learning approach

Lung adenocarcinoma (LUAD) is a tumour characterized by high tumour heterogeneity. Although there are numerous prognostic and immunotherapeutic options available for LUAD, there is a dearth of precise, individualized treatment plans. We integrated mRNA, lncRNA, microRNA, methylation and mutation data from the TCGA database for LUAD. Utilizing ten clustering algorithms, we identified stable multi-omics consensus clusters (MOCs). These data were then amalgamated with ten machine learning approaches to develop a robust model capable of reliably identifying patient prognosis and predicting immunotherapy outcomes. Through ten clustering algorithms, two prognostically relevant MOCs were identified, with MOC2 showing more favourable outcomes. We subsequently constructed a MOCs-associated machine learning model (MOCM) based on eight MOCs-specific hub genes. Patients characterized by a lower MOCM score exhibited better overall survival and responses to immunotherapy. These findings were consistent across multiple datasets, and compared to many previously published LUAD biomarkers, our MOCM score demonstrated superior predictive performance. Notably, the low MOCM group was more inclined towards 'hot' tumours, characterized by higher levels of immune cell infiltration. Intriguingly, a significant positive correlation between GJB3 and the MOCM score (R = 0.77, p < 0.01) was discovered. Further experiments confirmed that GJB3 significantly enhances LUAD proliferation, invasion and migration, indicating its potential as a key target for LUAD treatment. Our developed MOCM score accurately predicts the prognosis of LUAD patients and identifies potential beneficiaries of immunotherapy, offering broad clinical applicability.

Integrated multi-omics analysis and machine learning to refine molecular subtypes, prognosis, and immunotherapy in lung adenocarcinoma

Lung adenocarcinoma (LUAD) has a malignant characteristic that is highly aggressive and prone to metastasis. There is still a lack of suitable biomarkers to facilitate the refinement of precision-based therapeutic regimens. We used a combination of 10 known clustering algorithms and the omics data from 4 dimensions to identify high-resolution molecular subtypes of LUAD. Subsequently, consensus machine learning-related prognostic signature (CMRS) was developed based on subtypes related genes and an integrated program framework containing 10 machine learning algorithms. The efficiency of CMRS was analyzed from the perspectives of tumor microenvironment, genomic landscape, immunotherapy, drug sensitivity, and single-cell analysis. In terms of results, through multi-omics clustering, we identified 2 comprehensive omics subtypes (CSs) in which CS1 patients had worse survival outcomes, higher aggressiveness, mRNAsi and mutation frequency. Subsequently, we developed CMRS based on 13 key genes up-regulated in CS1. The prognostic predictive efficiency of CMRS was superior to most established LUAD prognostic signatures. CMRS demonstrated a strong correlation with tumor microenvironmental feature variants and genomic instability generation. Regarding clinical performance, patients in the high CMRS group were more likely to benefit from immunotherapy, whereas low CMRS were more likely to benefit from chemotherapy and targeted drug therapy. In addition, we evaluated that drugs such as neratinib, oligomycin A, and others may be candidates for patients in the high CMRS group. Single-cell analysis revealed that CMRS-related genes were mainly expressed in epithelial cells. The novel molecular subtypes identified in this study based on multi-omics data could provide new insights into the stratified treatment of LUAD, while the development of CMRS could serve as a candidate indicator of the degree of benefit of precision therapy and immunotherapy for LUAD.

Multi Omics Applications in Biological Systems

Traditional methodologies often fall short in addressing the complexity of biological systems. In this regard, system biology omics have brought invaluable tools for conducting comprehensive analysis. Current sequencing capabilities have revolutionized genetics and genomics studies, as well as the characterization of transcriptional profiling and dynamics of several species and sample types. Biological systems experience complex biochemical processes involving thousands of molecules. These processes occur at different levels that can be studied using mass spectrometry-based (MS-based) analysis, enabling high-throughput proteomics, glycoproteomics, glycomics, metabolomics, and lipidomics analysis. Here, we present the most up-to-date techniques utilized in the completion of omics analysis. Additionally, we include some interesting examples of the applicability of multi omics to a variety of biological systems.

Computational frameworks integrating deep learning and statistical models in mining multimodal omics data

BACKGROUND

In health research, multimodal omics data analysis is widely used to address important clinical and biological questions. Traditional statistical methods rely on the strong assumptions of distribution. Statistical methods such as testing and differential expression are commonly used in omics analysis. Deep learning, on the other hand, is an advanced computer science technique that is powerful in mining high-dimensional omics data for prediction tasks. Recently, integrative frameworks or methods have been developed for omics studies that combine statistical models and deep learning algorithms.

METHODS AND RESULTS

The aim of these integrative frameworks is to combine the strengths of both statistical methods and deep learning algorithms to improve prediction accuracy while also providing interpretability and explainability. This review report discusses the current state-of-the-art integrative frameworks, their limitations, and potential future directions in survival and time-to-event longitudinal analysis, dimension reduction and clustering, regression and classification, feature selection, and causal and transfer learning.

Integrated multiomics analysis and machine learning refine molecular subtypes and prognosis for muscle-invasive urothelial cancer

Muscle-invasive urothelial cancer (MUC), characterized by high aggressiveness and significant heterogeneity, is currently lacking highly precise individualized treatment options. We used a computational pipeline to synthesize multiomics data from MUC patients using 10 clustering algorithms, which were then combined with 10 machine learning algorithms to identify molecular subgroups of high resolution and develop a robust consensus machine learning-driven signature (CMLS). Through multiomics clustering, we identified three cancer subtypes (CSs) that are related to prognosis, with CS2 exhibiting the most favorable prognostic outcome. Subsequent screening enabled identification of 12 hub genes that constitute a CMLS with robust predictive power for prognosis. The low-CMLS group exhibited a more favorable prognosis and greater responsiveness to immunotherapy and was more likely to exhibit the "hot tumor" phenotype. The high-CMLS group had a poor prognosis and lower likelihood of benefitting from immunotherapy, but dasatinib and romidepsin may serve as promising treatments for them. Comprehensive analysis of multiomics data can offer important insights and further refine the molecular classification of MUC. Identification of CMLS represents a valuable tool for early prediction of patient prognosis and for screening potential candidates likely to benefit from immunotherapy, with broad implications for clinical practice.