Biology subjects

Azuaje, F.

Publications and source records attributed to Azuaje, F..

8 recordsLinked to original sources

bioRxiv2022

DrDimont: Explainable drug response predictionfrom differential analysis of multi-omics networks

MotivationWhile it has been well established that drugs affect and help patients differently, personalized drug response predictions remain challenging. Solutions based on single omics measurements have been proposed, and networks provide means to incorporate molecular interactions into reasoning. However, how to integrate the wealth of information contained in multiple omics layers still poses a complex problem. ResultsWe present DrDimont, Drug response prediction from Differential analysis of multi-omics networks. It allows for comparative conclusions between two conditions and translates them into differential drug response predictions. DrDimont focuses on molecular interactions. It establishes condition-specific networks from correlation within an omics layer that are then reduced and combined into heterogeneous, multi-omics molecular networks. A novel semi-local, path-based integration step ensures integrative conclusions. Differential predictions are derived from comparing the condition-specific integrated networks. DrDimonts predictions are explainable, i.e., molecular differences that are the source of high differential drug scores can be retrieved. We predict differential drug response in breast cancer using transcriptomics, proteomics, phosphosite, and metabolomics measurements and contrast estrogen receptor positive and receptor negative patients. DrDimont performs better than drug prediction based on differential protein expression or PageRank when evaluating it on ground truth data from cancer cell lines. We find proteomic and phosphosite layers to carry most information for distinguishing drug response. AvailabilityDrDimont is available on CRAN: https://cran.r-project.org/package=DrDimont. Contactkatharina.baum@hpi.de

bioinformatics↗

bioRxiv2022

BioNE: Integration of network embeddings for supervised learning

A network embedding approach reduces the analysis complexity of large biological networks by converting them to lowdimensional vector representations (features/embeddings). These lower-dimensional vectors can then be used in machine learning prediction tasks with a wide range of applications in computational biology and bioinformatics. Several network embedding approaches have been proposed with different methods of generating vector representations. These network embedding approaches can be quite diverse in terms of data representation and implementation. Moreover, most were not originally developed for biological networks. Therefore comparing and assessing the performance of these diverse models in practice, in biological contexts, can be challenging. To facilitate such comparisons, we have developed the BioNE framework for integration of different embedding methods in prediction tasks. Using this framework one can easily assess, for instance, whether combined vector representations from multiple embedding methods offer complementary information with regards to the network features and thus better performance on prediction tasks. In this paper, we present the BioNE software suite for embedding integration, which applies network embedding methods following standardised network preparation steps, and integrates the vector representations achieved by these methods using three different techniques. BioNE enables selection of prediction models, oversampling methods, feature selection methods, cross-validation type and cross-validation parameters. Availability and implementationBioNE pipeline and detailed explanation of implementation is freely available on GitHub, at https://github.com/pooryaparvizi/BioNE

bioinformatics↗

bioRxiv2020

Primary and recurrent glioma patient-derived orthotopic xenografts (PDOX) represent relevant patient avatars for precision medicine

Patient-derived cancer models are essential tools for studying tumor biology and preclinical interventions. Here, we show that glioma patient-derived orthotopic xenografts (PDOXs) enable long-term propagation of patient tumors and represent clinically relevant patient avatars. We created a large collection of PDOXs from primary and recurrent gliomas with and without mutations in IDH1, which retained histopathological, genetic, epigenetic and transcriptomic features of patient tumors with no mouse-specific clonal evolution. Longitudinal PDOX models recapitulate the limited genetic evolution of gliomas observed in patient tumors following treatment. PDOX-derived standardized tumor organoid cultures enabled assessment of drug responses, which were validated in mice. PDOXs showed clinically relevant responses to Temozolomide and to targeted treatments such as EGFR and CDK4/6 inhibitors in (epi)genetically defined groups, according to MGMT promoter and EGFR/CDK status respectively. Dianhydrogalactitol, a bifunctional alkylating agent, showed promising potential against glioblastoma. Our study underlines the clinical relevance of glioma PDOX models for translational research and personalized treatment studies.

cancer biology↗

bioRxiv2019

Machine learning-assisted neurotoxicity prediction in human midbrain organoids

A major challenge in the field of neurodegenerative diseases is the poor translation of pre-clinical models to clinical applications. The human brain is an immensely complex structure, which makes it difficult to recapitulate its development, function and disorders. In the recent years, brain organoids derived from human induced pluripotent stem cells have risen as novel tools to study neurodegenerative diseases such as Parkinsons disease (PD). PD is a multifactorial disorder, with aging, genetics and environmental factors as key etiological elements. The majority of the PD cases are idiopathic and proposed to result from a complex interaction between genetic predisposition and environmental exposure. Consequently, the identification of potentially disease causing environmental factors is of critical importance. Organoids, as complex multi-cellular tissue proxies, are an ideal tool to study cellular response to environmental changes. However, with increasing complexity of the system, usage of quantitative tools becomes challenging. This led us to develop an automated high-content image analysis pipeline for image-based cell profiling in the organoid system. Here, we introduce a midbrain organoid system that recapitulates features of neurotoxin-induced PD, representing a platform for machine-learning-assisted prediction of neurotoxicity in high-content imaging data. This model is a valuable tool for advanced in vitro PD modeling and for the screening of putative neurotoxic compounds.

pharmacology and toxicology↗

bioRxiv2019

Connecting Histopathology Imaging and Proteomics in Kidney Cancer through Machine Learning

Proteomics data encode molecular features of diagnostic value and accurately reflect key underlying biological mechanisms in cancers. Histopathology imaging is a well-established clinical approach to cancer diagnosis. The predictive relationship between large-scale proteomics and H&E-stained histopathology images remains largely uncharacterized. Here we investigate such associations through the application of machine learning, including deep neural networks, to proteomics and histology imaging datasets generated by the Clinical Proteomic Tumor Analysis Consortium (CPTAC) from clear cell renal cell carcinoma patients. We report robust correlations between a set of diagnostic proteins and predictions generated by an imaging-based classification model. Proteins significantly correlated with the histology-based predictions are significantly implicated in immune responses, extracellular matrix reorganization and metabolism. Moreover, we showed that the genes encoding these proteins also reliably recapitulate the biological associations with imaging-derived predictions based on strong gene-protein expression correlations. Our findings offer novel insights into the integrative modeling of histology and omics data through machine learning, as well as the methodological basis for new research opportunities in this and other cancer types.

systems biology↗

bioRxiv2019

A deep neural network approach to predicting clinical outcomes of neuroblastoma patients

The availability of high-throughput omics datasets from large patient cohorts has allowed the development of methods that aim at predicting patient clinical outcomes, such as survival and disease recurrence. Such methods are also important to better understand the biological mechanisms underlying disease etiology and development, as well as treatment responses. Recently, different predictive models, relying on distinct algorithms (including Support Vector Machines and Random Forests) have been investigated. In this context, deep learning strategies are of special interest due to their demonstrated superior performance over a wide range of problems and datasets. One of the main challenges of such strategies is the "small n large p" problem. Indeed, omics datasets typically consist of small numbers of samples and large numbers of features relative to typical deep learning datasets. Neural networks usually tackle this problem through feature selection or by including additional constraints during the learning process. We propose to tackle this problem with a novel strategy that relies on a graph-based method for feature extraction, coupled with a deep neural network for clinical outcome prediction. The omics data are first represented as graphs whose nodes represent patients, and edges represent correlations between the patients omics profiles. Topological features, such as centralities, are then extracted from these graphs for every node. Lastly, these features are used as input to train and test various classifiers. We apply this strategy to four neuroblastoma datasets and observe that models based on neural networks are more accurate than state of the art models (DNN: 85%-87%, SVM/RF: 75%-82%). We explore how different parameters and configurations are selected in order to overcome the effects of the small data problem as well as the curse of dimensionality. Our results indicate that the deep neural networks capture complex features in the data that help predicting patient clinical outcomes.

bioinformatics↗

bioRxiv2018

Independent component analysis provides clinically relevant insights into the biology of melanoma patients

The integration of publicly available and new patient-derived transcriptomic datasets is not straightforward and requires specialized approaches to deal with heterogeneity at technical and biological levels. Here we present a methodology that can overcome technical biases, predict clinically relevant outcomes and identify tumour-related biological processes in patients using previously collected large reference datasets. The approach is based on independent component analysis (ICA) - an unsupervised method of signal deconvolution. We developed parallel consensus ICA that robustly decomposes merged new and reference datasets into signals with minimal mutual dependency. By applying the method to a small cohort of primary melanoma and control samples combined with a large public melanoma dataset, we demonstrate that our method distinguishes cell-type specific signals from technical biases and allows to predict clinically relevant patient characteristics. Cancer subtypes, patient survival and activity of key tumour-related processes such as immune response, angiogenesis and cell proliferation were characterized. Additionally, through integration of transcriptomes and miRNomes, the method identified biological functions of miRNAs, which would otherwise not be possible.

genomics↗

bioRxiv2017

Dr.Paso: Drug response prediction and analysis system for oncology research

The prediction of anticancer drug response is crucial for achieving a more effective and precise treatment of patients. Models based on the analysis of large cell line collections have shown potential for investigating drug efficacy in a clinically-meaningful, cost-effective manner. Using data from thousands of cancer cell lines and drug response experiments, we propose a drug sensitivity prediction system based on a 47-gene expression profile, which was derived from an unbiased transcriptomic network analysis approach. The profile reflects the molecular activity of a diverse range of cancer-relevant processes and pathways. We validated our model using independent datasets and comparisons with published models. A high concordance between predicted and observed drug sensitivities was obtained, including additional validated predictions for four glioblastoma cell lines and four drugs. Our approach can accurately predict anti-cancer drug sensitivity and will enable further pre-clinical research. In the longer-term, it may benefit patient-oriented investigations and interventions.

bioinformatics↗