Identifying common transcriptome signatures of cancer by interpreting deep learning models

Interpretable deep cross networks unveiled common signatures of dysregulated epitranscriptomes across 12 cancer types

Cancer is a complex and multifaceted group of diseases characterized by uncontrolled cell growth that leads to the formation of malignant tumors. Recent studies suggest that N6-methyladenosine (m6A) RNA methylation plays pivotal roles in cancer pathology by influencing various cellular processes. However, the degree to which these mechanisms are shared across different cancer types remains unclear. In this study, we analyze an expansive array of 167 m6A epitranscriptome profiles covering 12 distinct cancer types and their originating normal tissues. We trained 12 distinct, cancer type-specific interpretable deep cross network models, which successfully distinguish between specific pairs of normal and cancer m6A contexts using integrated information from both the sequences and curated genomic knowledge. Interestingly, cross-cancer type testing indicated the existence of shared genomic patterns across various cancers at the epitranscriptome level. A pan-cancer model was subsequently developed to identify these shared patterns that could not be observed in a single cancer type. Our analysis uncovered, for the first time, a common epitranscriptome signature shared across multiple cancer types, particularly associated with RNA hybridization process and aberrant splicing. This highlights the importance of a comprehensive understanding of the pan-cancer epitranscriptome and holding potential implications in the development of RNA methylation-based therapeutics for various cancers.

Multimodal data fusion AI model uncovers tumor microenvironment immunotyping heterogeneity and enhanced risk stratification of breast cancer

Breast cancer is the leading cancer among women, with a significant number experiencing recurrence and metastasis, thereby reducing survival rates. This study focuses on the role of long noncoding RNAs (lncRNAs) in breast cancer immunotherapy response. We conducted an analysis involving 1027 patients from Sun Yat-sen Memorial Hospital, Sun Yat-sen University, and The Cancer Genome Atlas, utilizing RNA sequencing and pathology whole-slide images. We employed unsupervised clustering to identify distinct lncRNA expression patterns and developed an AI-based pathology model using convolutional neural networks to predict immune-metabolic subtypes. Additionally, we created a multimodal model integrating lncRNA data, immune-cell scores, clinical information, and pathology images for prognostic prediction. Our findings revealed four unique immune-metabolic subtypes, and the AI model demonstrated high predictive accuracy, highlighting the significant impact of lncRNAs on antitumor immunity and metabolic states within the tumor microenvironment. The AI-based pathology model, DeepClinMed-IM, exhibited high accuracy in predicting these subtypes. Additionally, the multimodal model, DeepClinMed-PGM, integrating pathology images, lncRNA data, immune-cell scores, and clinical information, showed superior prognostic performance. In conclusion, these AI models provide a robust foundation for precise prognostication and the identification of potential candidates for immunotherapy, advancing breast cancer research and treatment strategies.

Gene choice in cancer cells is exclusive in ion transport but concurrent in DNA replication

Cancers share common cellular and physiological features. Little is known about whether distinctive gene expression patterns can be displayed at the single-cell level by gene families in cancer cells. The expression of gene homologs within a family can exhibit concurrence and exclusivity. Concurrence can promote all-or-none expression patterns of related genes and underlie alternative physiological states. Conversely, exclusive gene families express the same or similar number of homologs in each cell, allowing a broad repertoire of cell identities to be generated. We show that gene families involved in the cell-cycle and antigen presentation are expressed concurrently. Concurrence in the DNA replication complex MCM reflects the replicative status of cells, including cell lines and cancer-derived organoids. Exclusive expression requires precise regulatory mechanism, but cancer cells retain this form of control for ion homeostasis and extend it to gene families involved in cell migration. Thus, the cell adhesion-based identity of healthy cells is transformed to an identity based on migration in the population of cancer cells, reminiscent of epithelial-mesenchymal transition.

Human lncRNAs harbor conserved modules embedded in different sequence contexts

We analyzed the structure of human long non-coding RNA (lncRNAs) genes to investigate whether the non-coding transcriptome is organized in modular domains, as is the case for protein-coding genes. To this aim, we compared all known human lncRNA exons and identified 340 pairs of exons with high sequence and/or secondary structure similarity but embedded in a dissimilar sequence context. We grouped these pairs in 106 clusters based on their reciprocal similarities. These shared modules are highly conserved between humans and the four great ape species, display evidence of purifying selection and likely arose as a result of recent segmental duplications. Our analysis contributes to the understanding of the mechanisms driving the evolution of the non-coding genome and suggests additional strategies towards deciphering the functional complexity of this class of molecules.

Genomic strategies for drug repurposing

Functional genomics, a multidisciplinary subject, investigates the functions of genes and their products in biological systems to better understand diseases and find new drugs. Drug repurposing is an economically efficient approach that entails discovering novel therapeutic applications for already-available medications. Genomics enables the identification of illness and therapeutic molecular characteristics and interactions, which in turn facilitates the process of drug repurposing. Techniques like gene expression profiling and Mendelian randomization are helpful in identifying possible medication candidates. Progress in computer science allows for the investigation and modeling of gene expression networks that involve large amounts of data. The amalgamation of data concerning DNA, RNA, and protein functions bears similarity to pharmacogenomics, a crucial aspect in crafting cancer therapeutics. Functional genomics in drug discovery, particularly for cancer, is still not thoroughly investigated, despite the existence of a significant amount of literature on the subject. Next-generation sequencing and proteomics present highly intriguing opportunities. Publicly available databases and mining techniques facilitate the development of cancer treatments based on functional genomics. Broadening the exploration and utilization of functional genomics holds significant potential for advancing drug discovery and repurposing, particularly within the realm of oncology.

Advanced Prediction of Hepatic Oncogenic Transformation in HBV Patients via RNA-Seq Data Analysis and Deep Learning Techniques

Liver cancer, recognized as a significant global health issue, is increasingly correlated with Hepatitis B virus (HBV) infection, as evidenced by numerous scientific studies. This study aims to examine the correlation between HBV infection and the development of liver cancer, focusing on using RNA sequencing (RNA-seq) to detect HBV sequences and applying deep learning techniques to estimate the likelihood of oncogenic transformation in individuals with HBV. Our study utilized RNA-seq data and employed Pathseq software and sophisticated deep learning models, including a convolutional neural network (CNN), to analyze the prevalence of HBV sequences in the samples of patients with liver cancer. Our research successfully identified the prevalence of HBV sequences and demonstrated that the CNN model achieved an exceptional Area Under the Curve (AUC) of 0.998 in predicting cancerous transformations. We observed no viral synergism that enhanced the pathogenicity of HBV. A detailed analysis of sequences misclassified by the CNN model revealed that longer sequences were more conducive to accurate recognition. The findings from this study provide critical insights into the management and prognosis of patients infected with HBV, highlighting the potential of advanced analytical techniques in understanding the complex interactions between viral infections and cancer development.

A Comparison of Tools That Identify Tumor Cells by Inferring Copy Number Variations from Single-Cell Experiments in Pancreatic Ductal Adenocarcinoma

Single-cell RNA sequencing (scRNA-seq) technique has enabled detailed analysis of gene expression at the single cell level, enhancing the understanding of subtle mechanisms that underly pathologies and drug resistance. To derive such biological meaning from sequencing data in oncology, some critical processing must be performed, including identification of the tumor cells by markers and algorithms that infer copy number variations (CNVs). We compared the performance of sciCNV, InferCNV, CopyKAT and SCEVAN tools that identify tumor cells by inferring CNVs from scRNA-seq data. Sequencing data from Pancreatic Ductal Adenocarcinoma (PDAC) patients, adjacent and healthy tissues were analyzed, and the predicted tumor cells were compared to those identified by well-assessed PDAC markers. Results from InferCNV, CopyKAT and SCEVAN overlapped by less than 30% with InferCNV showing the highest sensitivity (0.72) and SCEVAN the highest specificity (0.75). We show that the predictions are highly dependent on the sample and the software used, and that they return so many false positives hence are of little use in verifying or filtering predictions made via tumor biomarkers. We highlight how critical this processing can be, warn against the blind use of these software and point out the great need for more reliable algorithms.

Designing interpretable deep learning applications for functional genomics: a quantitative analysis

Deep learning applications have had a profound impact on many scientific fields, including functional genomics. Deep learning models can learn complex interactions between and within omics data; however, interpreting and explaining these models can be challenging. Interpretability is essential not only to help progress our understanding of the biological mechanisms underlying traits and diseases but also for establishing trust in these model's efficacy for healthcare applications. Recognizing this importance, recent years have seen the development of numerous diverse interpretability strategies, making it increasingly difficult to navigate the field. In this review, we present a quantitative analysis of the challenges arising when designing interpretable deep learning solutions in functional genomics. We explore design choices related to the characteristics of genomics data, the neural network architectures applied, and strategies for interpretation. By quantifying the current state of the field with a predefined set of criteria, we find the most frequent solutions, highlight exceptional examples, and identify unexplored opportunities for developing interpretable deep learning models in genomics.

A comprehensive survey on deep learning-based identification and predicting the interaction mechanism of long non-coding RNAs

Long noncoding RNAs (lncRNAs) have been discovered to be extensively involved in eukaryotic epigenetic, transcriptional, and post-transcriptional regulatory processes with the advancements in sequencing technology and genomics research. Therefore, they play crucial roles in the body's normal physiology and various disease outcomes. Presently, numerous unknown lncRNA sequencing data require exploration. Establishing deep learning-based prediction models for lncRNAs provides valuable insights for researchers, substantially reducing time and costs associated with trial and error and facilitating the disease-relevant lncRNA identification for prognosis analysis and targeted drug development as the era of artificial intelligence progresses. However, most lncRNA-related researchers lack awareness of the latest advancements in deep learning models and model selection and application in functional research on lncRNAs. Thus, we elucidate the concept of deep learning models, explore several prevalent deep learning algorithms and their data preferences, conduct a comprehensive review of recent literature studies with exemplary predictive performance over the past 5 years in conjunction with diverse prediction functions, critically analyze and discuss the merits and limitations of current deep learning models and solutions, while also proposing prospects based on cutting-edge advancements in lncRNA research.

Multi-omics based artificial intelligence for cancer research

With significant advancements of next generation sequencing technologies, large amounts of multi-omics data, including genomics, epigenomics, transcriptomics, proteomics, and metabolomics, have been accumulated, offering an unprecedented opportunity to explore the heterogeneity and complexity of cancer across various molecular levels and scales. One of the promising aspects of multi-omics lies in its capacity to offer a holistic view of the biological networks and pathways underpinning cancer, facilitating a deeper understanding of its development, progression, and response to treatment. However, the exponential growth of data generated by multi-omics studies present significant analytical challenges. Processing, analyzing, integrating, and interpreting these multi-omics datasets to extract meaningful insights is an ambitious task that stands at the forefront of current cancer research. The application of artificial intelligence (AI) has emerged as a powerful solution to these challenges, demonstrating exceptional capabilities in deciphering complex patterns and extracting valuable information from large-scale, intricate omics datasets. This review delves into the synergy of AI and multi-omics, highlighting its revolutionary impact on oncology. We dissect how this confluence is reshaping the landscape of cancer research and clinical practice, particularly in the realms of early detection, diagnosis, prognosis, treatment and pathology. Additionally, we elaborate the latest AI methods for multi-omics integration to provide a comprehensive insight of the complex biological mechanisms and inherent heterogeneity of cancer. Finally, we discuss the current challenges of data harmonization, algorithm interpretability, and ethical considerations. Addressing these challenges necessitates a multidisciplinary collaboration, paving the promising way for more precise, personalized, and effective treatments for cancer patients.

Pathologic features of brain hemorrhage after radiation treatment: case series with somatic mutation analysis

BACKGROUND

Radiation treatment for diseases of the brain can result in hemorrhagic adverse radiation effects. The underlying pathologic substrate of brain bleeding after irradiation has not been elucidated, nor potential associations with induced somatic mutations.

METHODS

We retrospectively reviewed our department's pathology database over 5 years and identified 5 biopsy specimens (4 patients) for hemorrhagic lesions after brain irradiation. Tissues with active malignancy were excluded. Samples were characterized using H&E, Perl's Prussian Blue, and Masson's Trichrome; immunostaining for B-cells (anti-CD20), T-cells (anti-CD3), endothelium (anti-CD31), macrophages (anti-CD163), α-smooth muscle actin, and TUNEL. DNA analysis was done by two panels of next-generation sequencing for somatic mutations associated with known cerebrovascular anomalies.

RESULTS

One lesion involved hemorrhagic expansion among multifocal microbleeds that had developed after craniospinal irradiation for distant medulloblastoma treatment. Three bleeds arose in the bed of focally irradiated arteriovenous malformations (AVM) after confirmed obliteration. A fifth specimen involved the radiation field distinct from an irradiated AVM bed. From these, 2 patterns of hemorrhagic vascular pathology were identified: encapsulated hematomas and cavernous-like malformations. All lesions included telangiectasias with dysmorphic endothelium, consistent with primordial cavernous malformations with an associated inflammatory response. DNA analysis demonstrated genetic variants in PIK3CA and/or PTEN genes but excluded mutations in CCM genes.

CONCLUSIONS

Despite pathologic heterogeneity, brain bleeding after irradiation is uniformly associated with primordial cavernous-like telangiectasias and disruption of genes implicated in dysangiogenesis but not genes implicated as causative of cerebral cavernous malformations. This may implicate a novel signaling axis as an area for future study.

Genomic Profiling and Molecular Characterisation of Metastatic Urothelial Carcinoma

The clinical management of metastatic urothelial carcinoma (mUC) is undergoing a major paradigm shift; the integration of immune checkpoint inhibitors (ICIs) and antibody-drug conjugates (ADCs) into the mUC therapeutic strategy has succeeded in improving platinum-based chemotherapy outcomes. Given the expanding therapeutic armamentarium, it is crucial to identify efficacy-predictive biomarkers that can guide an individual patient's therapeutic strategy. We reviewed the literature data on mUC genomic alterations of clinical interest, discussing their prognostic and predictive role. In particular, we explored the role of the fibroblast growth factor receptor (FGFR) family, epidermal growth factor receptor 2 (HER2), mechanistic target of rapamycin (mTOR) axis, DNA repair genes, and microsatellite instability. Currently, based on the available clinical data, FGFR inhibitors and HER2-directed ADCs are effective therapeutic options for later lines of biomarker-driven mUC. However, emerging genomic data highlight the opportunity for earlier use and/or combination with other drugs of both FGFR inhibitors and HER2-directed ADCs and also reveal additional potential drug targets that could change mUC management.

MAJIQlopedia: an encyclopedia of RNA splicing variations in human tissues and cancer

Quantification of RNA splicing variations based on RNA-Sequencing can reveal tissue- and disease-specific splicing patterns. To study such splicing variations, we introduce MAJIQlopedia, an encyclopedia of splicing variations that encompasses 86 human tissues and 41 cancer datasets. MAJIQlopedia reports annotated and unannotated splicing events for a total of 486 175 alternative splice junctions in normal tissues and 338 317 alternative splice junctions in cancer. This database, available at https://majiq.biociphers.org/majiqlopedia/, includes a user-friendly interface that provides graphical representations of junction usage quantification for each junction across all tissue or cancer types. To demonstrate case usage of MAJIQlopedia, we review splicing variations in genes WT1, MAPT and BIN1, which all have known tissue or cancer-specific splicing variations. We also use MAJIQlopedia to highlight novel splicing variations in FDX1 and MEGF9 in normal tissues, and we uncover a novel exon inclusion event in RPS6KA6 that only occurs in two cancer types. Users can download the database, request the addition of data to the webtool, or install a MAJIQlopedia server to integrate proprietary data. MAJIQlopedia can serve as a reference database for researchers seeking to understand what splicing variations exist in genes of interest, and those looking to understand tissue- or cancer-specific splice isoform usage.

Immune, metabolic landscapes of prognostic signatures for lung adenocarcinoma based on a novel deep learning framework

Lung adenocarcinoma (LUAD) is a malignant tumor with high lethality, and the aim of this study was to identify promising biomarkers for LUAD. Using the TCGA-LUAD dataset as a discovery cohort, a novel joint framework VAEjMLP based on variational autoencoder (VAE) and multilayer perceptron (MLP) was proposed. And the Shapley Additive Explanations (SHAP) method was introduced to evaluate the contribution of feature genes to the classification decision, which helped us to develop a biologically meaningful biomarker potential scoring algorithm. Nineteen potential biomarkers for LUAD were identified, which were involved in the regulation of immune and metabolic functions in LUAD. A prognostic risk model for LUAD was constructed by the biomarkers HLA-DRB1, SCGB1A1, and HLA-DRB5 screened by Cox regression analysis, dividing the patients into high-risk and low-risk groups. The prognostic risk model was validated with external datasets. The low-risk group was characterized by enrichment of immune pathways and higher immune infiltration compared to the high-risk group. While, the high-risk group was accompanied by an increase in metabolic pathway activity. There were significant differences between the high- and low-risk groups in metabolic reprogramming of aerobic glycolysis, amino acids, and lipids, as well as in angiogenic activity, epithelial-mesenchymal transition, tumorigenic cytokines, and inflammatory response. Furthermore, high-risk patients were more sensitive to Afatinib, Gefitinib, and Gemcitabine as predicted by the pRRophetic algorithm. This study provides prognostic signatures capable of revealing the immune and metabolic landscapes for LUAD, and may shed light on the identification of other cancer biomarkers.

Novel research and future prospects of artificial intelligence in cancer diagnosis and treatment

Research into the potential benefits of artificial intelligence for comprehending the intricate biology of cancer has grown as a result of the widespread use of deep learning and machine learning in the healthcare sector and the availability of highly specialized cancer datasets. Here, we review new artificial intelligence approaches and how they are being used in oncology. We describe how artificial intelligence might be used in the detection, prognosis, and administration of cancer treatments and introduce the use of the latest large language models such as ChatGPT in oncology clinics. We highlight artificial intelligence applications for omics data types, and we offer perspectives on how the various data types might be combined to create decision-support tools. We also evaluate the present constraints and challenges to applying artificial intelligence in precision oncology. Finally, we discuss how current challenges may be surmounted to make artificial intelligence useful in clinical settings in the future.

Attention-based generative adversarial networks improve prognostic outcome prediction of cancer from multimodal data

The prediction of prognostic outcome is critical for the development of efficient cancer therapeutics and potential personalized medicine. However, due to the heterogeneity and diversity of multimodal data of cancer, data integration and feature selection remain a challenge for prognostic outcome prediction. We proposed a deep learning method with generative adversarial network based on sequential channel-spatial attention modules (CSAM-GAN), a multimodal data integration and feature selection approach, for accomplishing prognostic stratification tasks in cancer. Sequential channel-spatial attention modules equipped with an encoder-decoder are applied for the input features of multimodal data to accurately refine selected features. A discriminator network was proposed to make the generator and discriminator learning in an adversarial way to accurately describe the complex heterogeneous information of multiple modal data. We conducted extensive experiments with various feature selection and classification methods and confirmed that the CSAM-GAN via the multilayer deep neural network (DNN) classifier outperformed these baseline methods on two different multimodal data sets with miRNA expression, mRNA expression and histopathological image data: lower-grade glioma and kidney renal clear cell carcinoma. The CSAM-GAN via the multilayer DNN classifier bridges the gap between heterogenous multimodal data and prognostic outcome prediction.

Deep Learning Techniques with Genomic Data in Cancer Prognosis: A Comprehensive Review of the 2021-2023 Literature

Deep learning has brought about a significant transformation in machine learning, leading to an array of novel methodologies and consequently broadening its influence. The application of deep learning in various sectors, especially biomedical data analysis, has initiated a period filled with noteworthy scientific developments. This trend has majorly influenced cancer prognosis, where the interpretation of genomic data for survival analysis has become a central research focus. The capacity of deep learning to decode intricate patterns embedded within high-dimensional genomic data has provoked a paradigm shift in our understanding of cancer survival. Given the swift progression in this field, there is an urgent need for a comprehensive review that focuses on the most influential studies from 2021 to 2023. This review, through its careful selection and thorough exploration of dominant trends and methodologies, strives to fulfill this need. The paper aims to enhance our existing understanding of applications of deep learning in cancer survival analysis, while also highlighting promising directions for future research. This paper undertakes aims to enrich our existing grasp of the application of deep learning in cancer survival analysis, while concurrently shedding light on promising directions for future research in this vibrant and rapidly proliferating field.

Deep Learning Approaches for lncRNA-Mediated Mechanisms: A Comprehensive Review of Recent Developments

This review paper provides an extensive analysis of the rapidly evolving convergence of deep learning and long non-coding RNAs (lncRNAs). Considering the recent advancements in deep learning and the increasing recognition of lncRNAs as crucial components in various biological processes, this review aims to offer a comprehensive examination of these intertwined research areas. The remarkable progress in deep learning necessitates thoroughly exploring its latest applications in the study of lncRNAs. Therefore, this review provides insights into the growing significance of incorporating deep learning methodologies to unravel the intricate roles of lncRNAs. By scrutinizing the most recent research spanning from 2021 to 2023, this paper provides a comprehensive understanding of how deep learning techniques are employed in investigating lncRNAs, thereby contributing valuable insights to this rapidly evolving field. The review is aimed at researchers and practitioners looking to integrate deep learning advancements into their lncRNA studies.

Emerging roles and potential clinical applications of long non-coding RNAs in hepatocellular carcinoma

Hepatocellular carcinoma is one of the most common highly malignant tumors in humans, as well as the leading cause of cancer-related death worldwide. Growing evidence has indicated that lncRNAs are implicated in different molecular mechanisms, including interactions with DNA, RNA, or protein, so that to regulate the gene expression at epigenetic, transcriptional, or posttranscriptional level. Moreover, the mechanism of action of lncRNA is closely related to its subcellular localization. An increasing number of studies have certified that lncRNA plays a significant biological function in the occurrence and development of hepatocellular carcinoma, such as involving in cell proliferation, metastasis, apoptosis, ferroptosis, autophagy, and reprogramming of energy metabolism. As a result, lncRNA has great potential as a novel biomarker for diagnosis or therapeutics of hepatocellular carcinoma. In this review, we highlight the correlation between subcellular localization of lncRNA and its mechanism of action, discuss the biological roles of lncRNA and the latest research advances in hepatocellular carcinoma, and emphasize the potential of lncRNA as a therapeutic target for advanced patients of hepatocellular carcinoma.