Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models.
Ontology highlight
ABSTRACT:
SUBMITTER: Yousefi S
PROVIDER: S-EPMC5601479 | biostudies-literature | 2017 Sep
REPOSITORIES: biostudies-literature
Publications
Predicting clinical outcomes from large scale cancer genomic profiles with deep survival models.
Yousefi Safoora S Amrollahi Fatemeh F Amgad Mohamed M Dong Chengliang C Lewis Joshua E JE Song Congzheng C Gutman David A DA Halani Sameer H SH Velazquez Vega Jose Enrique JE Brat Daniel J DJ Cooper Lee A D LAD
Scientific reports 20170915 1
Translating the vast data generated by genomic platforms into accurate predictions of clinical outcomes is a fundamental challenge in genomic medicine. Many prediction methods face limitations in learning from the high-dimensional profiles generated by these platforms, and rely on experts to hand-select a small number of features for training prediction models. In this paper, we demonstrate how deep learning and Bayesian optimization methods that have been remarkably successful in general high-d ...[more]
Similar Datasets
Project description:Polygenic risk scores (PRSs) are expected to play a critical role in precision medicine. Currently, PRS predictors are generally based on linear models using summary statistics, and more recently individual-level data. However, these predictors mainly capture additive relationships and are limited in data modalities they can use. We developed a deep learning framework (EIR) for PRS prediction which includes a model, genome-local-net (GLN), specifically designed for large-scale genomics data. The framework supports multi-task learning, automatic integration of other clinical and biochemical data, and model explainability. When applied to individual-level data from the UK Biobank, the GLN model demonstrated a competitive performance compared to established neural network architectures, particularly for certain traits, showcasing its potential in modeling complex genetic relationships. Furthermore, the GLN model outperformed linear PRS methods for Type 1 Diabetes, likely due to modeling non-additive genetic effects and epistasis. This was supported by our identification of widespread non-additive genetic effects and epistasis in the context of T1D. Finally, we constructed PRS models that integrated genotype, blood, urine, and anthropometric data and found that this improved performance for 93% of the 290 diseases and disorders considered. EIR is available at https://github.com/arnor-sigurdsson/EIR.
| S-EPMC10325897 | biostudies-literature
Project description:MotivationOne of the major goals in large-scale genomic studies is to identify genes with a prognostic impact on time-to-event outcomes which provide insight into the disease process. With rapid developments in high-throughput genomic technologies in the past two decades, the scientific community is able to monitor the expression levels of tens of thousands of genes and proteins resulting in enormous datasets where the number of genomic features is far greater than the number of subjects. Methods based on univariate Cox regression are often used to select genomic features related to survival outcome; however, the Cox model assumes proportional hazards (PH), which is unlikely to hold for each feature. When applied to genomic features exhibiting some form of non-proportional hazards (NPH), these methods could lead to an under- or over-estimation of the effects. We propose a broad array of marginal screening techniques that aid in feature ranking and selection by accommodating various forms of NPH. First, we develop an approach based on Kullback-Leibler information divergence and the Yang-Prentice model that includes methods for the PH and proportional odds (PO) models as special cases. Next, we propose R2 measures for the PH and PO models that can be interpreted in terms of explained randomness. Lastly, we propose a generalized pseudo-R2 index that includes PH, PO, crossing hazards and crossing odds models as special cases and can be interpreted as the percentage of separability between subjects experiencing the event and not experiencing the event according to feature measurements.ResultsWe evaluate the performance of our measures using extensive simulation studies and publicly available datasets in cancer genomics. We demonstrate that the proposed methods successfully address the issue of NPH in genomic feature selection and outperform existing methods.Availability and implementationR code for the proposed methods is available at github.com/lburns27/Feature-Selection.Contactkarthik.devarajan@fccc.edu.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC7267818 | biostudies-literature
Project description:BACKGROUND:The study of high-throughput genomic profiles from a pharmacogenomics viewpoint has provided unprecedented insights into the oncogenic features modulating drug response. A recent study screened for the response of a thousand human cancer cell lines to a wide collection of anti-cancer drugs and illuminated the link between cellular genotypes and vulnerability. However, due to essential differences between cell lines and tumors, to date the translation into predicting drug response in tumors remains challenging. Recently, advances in deep learning have revolutionized bioinformatics and introduced new techniques to the integration of genomic data. Its application on pharmacogenomics may fill the gap between genomics and drug response and improve the prediction of drug response in tumors. RESULTS:We proposed a deep learning model to predict drug response (DeepDR) based on mutation and expression profiles of a cancer cell or a tumor. The model contains three deep neural networks (DNNs), i) a mutation encoder pre-trained using a large pan-cancer dataset (The Cancer Genome Atlas; TCGA) to abstract core representations of high-dimension mutation data, ii) a pre-trained expression encoder, and iii) a drug response predictor network integrating the first two subnetworks. Given a pair of mutation and expression profiles, the model predicts IC50 values of 265 drugs. We trained and tested the model on a dataset of 622 cancer cell lines and achieved an overall prediction performance of mean squared error at 1.96 (log-scale IC50 values). The performance was superior in prediction error or stability than two classical methods (linear regression and support vector machine) and four analog DNN models of DeepDR, including DNNs built without TCGA pre-training, partly replaced by principal components, and built on individual types of input data. We then applied the model to predict drug response of 9059 tumors of 33 cancer types. Using per-cancer and pan-cancer settings, the model predicted both known, including EGFR inhibitors in non-small cell lung cancer and tamoxifen in ER+ breast cancer, and novel drug targets, such as vinorelbine for TTN-mutated tumors. The comprehensive analysis further revealed the molecular mechanisms underlying the resistance to a chemotherapeutic drug docetaxel in a pan-cancer setting and the anti-cancer potential of a novel agent, CX-5461, in treating gliomas and hematopoietic malignancies. CONCLUSIONS:Here we present, as far as we know, the first DNN model to translate pharmacogenomics features identified from in vitro drug screening to predict the response of tumors. The results covered both well-studied and novel mechanisms of drug resistance and drug targets. Our model and findings improve the prediction of drug response and the identification of novel therapeutic options.
| S-EPMC6357352 | biostudies-literature
Project description:Approximately 90% of pre-clinically validated drugs fail in clinical trials due to unanticipated clinical outcomes, costing over several hundred million US dollars per drug. Despite such critical importance, translating pre-clinical data to clinical outcomes remain a major challenge. Herein, we designed a modality-independent and unbiased approach to predict clinical outcomes of drugs. The approach exploits their multi-organ transcriptome patterns induced in mice and a unique mouse-transcriptome database “humanized” by machine learning algorithms and human clinical outcome datasets. The cross-validation with small-molecule, antibody and peptide drugs shows effective and efficient identification of the previously known outcomes of 5,519 adverse events and 11,312 therapeutic indications. In addition, the approach is adaptable to deducing potential molecular mechanisms underlying these outcomes. Furthermore, the approach identifies previously unsuspected repositioning targets. These results, together with the fact that it requires no prior structural or mechanistic information of drugs, illustrate its versatile applications to drug development process.
2020-01-09 | GSE142068 | GEO
Project description:Applications of generative models for genomic data have gained significant momentum in the past few years, with scopes ranging from data characterization to generation of genomic segments and functional sequences. In our previous study, we demonstrated that generative adversarial networks (GANs) and restricted Boltzmann machines (RBMs) can be used to create novel high-quality artificial genomes (AGs) which can preserve the complex characteristics of real genomes such as population structure, linkage disequilibrium and selection signals. However, a major drawback of these models is scalability, since the large feature space of genome-wide data increases computational complexity vastly. To address this issue, we implemented a novel convolutional Wasserstein GAN (WGAN) model along with a novel conditional RBM (CRBM) framework for generating AGs with high SNP number. These networks implicitly learn the varying landscape of haplotypic structure in order to capture complex correlation patterns along the genome and generate a wide diversity of plausible haplotypes. We performed comparative analyses to assess both the quality of these generated haplotypes and the amount of possible privacy leakage from the training data. As the importance of genetic privacy becomes more prevalent, the need for effective privacy protection measures for genomic data increases. We used generative neural networks to create large artificial genome segments which possess many characteristics of real genomes without substantial privacy leakage from the training dataset. In the near future, with further improvements in haplotype quality and privacy preservation, large-scale artificial genome databases can be assembled to provide easily accessible surrogates of real databases, allowing researchers to conduct studies with diverse genomic data within a safe ethical framework in terms of donor privacy.
| S-EPMC10635570 | biostudies-literature
Project description:BackgroundNeuroblastoma is a heterogeneous disease with diverse clinical outcomes. Current risk group models require improvement as patients within the same risk group can still show variable prognosis. Recently collected genome-wide datasets provide opportunities to infer neuroblastoma subtypes in a more unified way. Within this context, data integration is critical as different molecular characteristics can contain complementary signals. To this end, we utilized the genomic datasets available for the SEQC cohort patients to develop supervised and unsupervised models that can predict disease prognosis.ResultsOur supervised model trained on the SEQC cohort can accurately predict overall survival and event-free survival profiles of patients in two independent cohorts. We also performed extensive experiments to assess the prediction accuracy of high risk patients and patients without MYCN amplification. Our results from this part suggest that clinical endpoints can be predicted accurately across multiple cohorts. To explore the data in an unsupervised manner, we used an integrative clustering strategy named multi-view kernel k-means (MVKKM) that can effectively integrate multiple high-dimensional datasets with varying weights. We observed that integrating different gene expression datasets results in a better patient stratification compared to using these datasets individually. Also, our identified subgroups provide a better Cox regression model fit compared to the existing risk group definitions.ConclusionAltogether, our results indicate that integration of multiple genomic characterizations enables the discovery of subtypes that improve over existing definitions of risk groups. Effective prediction of survival times will have a direct impact on choosing the right therapies for patients.ReviewersThis article was reviewed by Susmita Datta, Wenzhong Xiao and Ziv Shkedy.
| S-EPMC6889397 | biostudies-literature
Project description:A tantalizing question in cellular physiology is whether the cellular state and environmental conditions can be inferred by the expression signature of an organism. To investigate this relationship, we created an extensive normalized gene expression compendium for the bacterium Escherichia coli that was further enriched with meta-information through an iterative learning procedure. We then constructed an ensemble method to predict environmental and cellular state, including strain, growth phase, medium, oxygen level, antibiotic and carbon source presence. Results show that gene expression is an excellent predictor of environmental structure, with multi-class ensemble models achieving balanced accuracy between 70.0% (±3.5%) to 98.3% (±2.3%) for the various characteristics. Interestingly, this performance can be significantly boosted when environmental and strain characteristics are simultaneously considered, as a composite classifier that captures the inter-dependencies of three characteristics (medium, phase and strain) achieved 10.6% (±1.0%) higher performance than any individual models. Contrary to expectations, only 59% of the top informative genes were also identified as differentially expressed under the respective conditions. Functional analysis of the respective genetic signatures implicates a wide spectrum of Gene Ontology terms and KEGG pathways with condition-specific information content, including iron transport, transferases, and enterobactin synthesis. Further experimental phenotypic-to-genotypic mapping that we conducted for knock-out mutants argues for the information content of top-ranked genes. This work demonstrates the degree at which genome-scale transcriptional information can be predictive of latent, heterogeneous and seemingly disparate phenotypic and environmental characteristics, with far-reaching applications.
| S-EPMC4361189 | biostudies-literature
Project description:MotivationThe interaction of miRNA and lncRNA is known to be important for gene regulations. However, not many computational approaches have been developed to analyze known interactions and predict the unknown ones. Given that there are now more evidences that suggest that lncRNA-miRNA interactions are closely related to their relative expression levels in the form of a titration mechanism, we analyzed the patterns in large-scale expression profiles of known lncRNA-miRNA interactions. From these uncovered patterns, we noticed that lncRNAs tend to interact collaboratively with miRNAs of similar expression profiles, and vice versa.ResultsBy representing known interaction between lncRNA and miRNA as a bipartite graph, we propose here a technique, called EPLMI, to construct a prediction model from such a graph. EPLMI performs its tasks based on the assumption that lncRNAs that are highly similar to each other tend to have similar interaction or non-interaction patterns with miRNAs and vice versa. The effectiveness of the prediction model so constructed has been evaluated using the latest dataset of lncRNA-miRNA interactions. The results show that the prediction model can achieve AUCs of 0.8522 and 0.8447 ± 0.0017 based on leave-one-out cross validation and 5-fold cross validation. Using this model, we show that lncRNA-miRNA interactions can be reliably predicted. We also show that we can use it to select the most likely lncRNA targets that specific miRNAs would interact with. We believe that the prediction models discovered by EPLMI can yield great insights for further research on ceRNA regulation network. To the best of our knowledge, EPLMI is the first technique that is developed for large-scale lncRNA-miRNA interaction profiling.Availability and implementationMatlab codes and dataset are available at https://github.com/yahuang1991polyu/EPLMI/.Contactyu-an.huang@connect.polyu.hk or zhuhongyou@ms.xjb.ac.cn.Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC6192210 | biostudies-literature
Project description:BackgroundIn recent years, breast cancer has become the most common cancer in the world, increasing women's health risks. Approximately 60% of breast cancers are categorized as human epidermal growth factor receptor 2 (HER2)-low tumors. Recently, antibody-drug conjugates have been found to have positive anticancer efficacy in patients with HER2-low breast cancer, but more studies are required to comprehend their clinical and molecular characteristics.MethodsIn this study, we retrospectively analyzed the data of 165 early breast cancer patients with pT1-2N1M0 who had undergone the RecurIndex testing. To better understand HER2-low tumors, we investigated the RecurIndex genomic profiles, clinicopathologic features, and survival outcomes of breast cancers according to HER2 status.ResultsFirst, there were significantly more hormone receptor (HR)-positive tumors, luminal-type tumors, and low Ki67 levels in the HER2-low than in the HER2-zero. Second, RI-LR (P = .0294) and RI-DR (P = .001) scores for HER2-low and HER2-zero were statistically significant. Third, within HER2-negative disease, HR-positive/HER2-low tumors showed highest ESR1, NFATC2IP, PTI1, ERBB2, and OBSL1 expressions. Fourth, results of the survival analysis showed that lower expression of HER2 was associated with improved relapse-free survival for HR-positive tumors, but not for HR-negative tumors.ConclusionsThe present study highlights the unique features of HER2-low tumors in terms of their clinical characteristics as well as their gene expression profiles. HR status may influence the prognosis of patients with HER2-low expression, and patients with HR-positive/HER2-low expression may have a favorable outcome.
| S-EPMC10712905 | biostudies-literature
Project description:BackgroundMetastases are responsible for over 70% of deaths from lung adenocarcinomas. Previous large-scale studies of LUAD mainly focused on primary diseases. We aimed to comprehensively analyze the genomic landscape of metastatic LUADs and elucidate its clinical implications in the context of precision medicine.MethodsWe performed retrospective analyses on targeted sequencing data of 3,743 primary tumors and 934 metastases from 4,480 patients with lung adenocarcinomas, and PD-L1 immunohistochemical data of 1,336 primary tumors and 252 metastases from 1,588 LUAD patients.ResultsMetastases generally manifested significantly higher mutational burdens and chromosomal instability than primary lung adenocarcinomas. Clinically actionable alterations, including ALK mutations, ALK and ROS1 fusions, and MET copy number gains, were enriched in metastases, particularly metastases to some specific organs/tissues, such as lymph nodes, liver, and brain. PD-L1 expression decreased as the approximate metastatic distance increased. Additional data of paired primary tumors and metastases to lymph nodes and brain validated patterns of actionable alterations and candidates for metastatic drivers. Two evolutionary modes of metastatic dissemination, common origins and distinct origins, were identified in both types of primary-metastasis pairs.ConclusionsOur study showed heterogenous patterns of clinically actionable alterations, PD-L1 expressions, metastatic driver candidates, and evolutionary patterns among multiple types of metastases of lung adenocarcinomas, which may advise the planning of treatments and the identification of novel therapeutic targets.
| S-EPMC8571538 | biostudies-literature