Related references
Note: Only part of the references are listed.
Article
Biochemical Research Methods
Daria Doncevic et al.
Summary: In this study, a novel VAE model called OntoVAE is designed to incorporate any ontology in its latent space and decoder part, providing pathway or phenotype activities for ontology terms. OntoVAE is demonstrated to be applicable in predictive modeling and is able to predict the effects of genetic or drug-induced perturbations using different ontologies and both bulk and single-cell transcriptomic datasets. Furthermore, a flexible framework is provided, which can be easily adapted to any ontology and dataset.
Article
Biochemical Research Methods
Magdalena Wysocka et al.
Summary: This systematic review discusses the use of Deep Learning (DL) models in oncology, focusing on multi-omics analysis. The study found that the current DL models have limited transparency and explainability in the biomedical domain. However, recent research has shown that by integrating prior biological knowledge, such as biological relational and network knowledge, the generalization and interpretability of DL models can be improved, which is crucial for the development of DL models with biological interpretability.
BMC BIOINFORMATICS
(2023)
Review
Genetics & Heredity
Gherman Novakovsky et al.
Summary: This Review describes the advances in explainable artificial intelligence (xAI) in genomics, where researchers are using deep learning approaches to gain biological insights behind the models, moving beyond the traditional 'black box' nature.
NATURE REVIEWS GENETICS
(2023)
Article
Multidisciplinary Sciences
Chirag Agarwal et al.
Summary: As the understanding of graph neural networks (GNNs) relies increasingly on explanations, it becomes crucial to evaluate the quality and reliability of these explanations. However, existing graph datasets lack or have unreliable ground-truth explanations, making it challenging to assess GNN explanations. To address this, we introduce ShapeGGen, a synthetic graph data generator capable of producing benchmark datasets with varying graph characteristics and accompanied by ground-truth explanations. By emulating diverse real-world scenarios, ShapeGGen offers flexibility in generating synthetic datasets and corresponding ground-truth explanations. These, along with real-world graph datasets, are included in GraphXAI, a graph explainability library that also provides data loaders, data processing functions, visualizers, GNN model implementations, and evaluation metrics for benchmarking GNN explainability methods.
Article
Biology
Pelin Gundogdu et al.
Summary: Single-cell data enables the study of cell dynamics at unprecedented resolution. SigPrimedNet is a data-driven solution that identifies cells and learns the functional summarization of signaling measurements.
Article
Biochemical Research Methods
Luigi Ferraro et al.
Summary: This study developed an artificial neural network model for predicting drug sensitivity, enhancing interpretability. The model utilized multiomics data from different tumor tissue sources and molecular descriptors encapsulating drug properties. Results showed favorable outcomes in predicting drug synergy, outperforming state-of-the-art machine learning algorithms in publicly available drug screening datasets.
Article
Biotechnology & Applied Microbiology
Joseph D. Janizek et al.
Summary: As interest in using unsupervised deep learning models for gene expression data analysis grows, more methods have been developed to enhance the interpretability of these models. These methods can be categorized into post hoc analyses of black box models using feature attribution methods, and approaches to construct inherently interpretable models with biologically-constrained architectures. In this study, the authors propose PAUSE, an unsupervised pathway attribution method that identifies major sources of transcriptomic variation when combined with biologically-constrained neural network models.
Article
Biochemical Research Methods
Victoria Bourgeais et al.
Summary: In this article, a knowledge-based deep learning model called GraphGONet is proposed, which incorporates Gene Ontology into the hidden layers to achieve a self-explaining neural network. The experiments confirm the accuracy and interpretability of the model, making it promising and valuable in the medical field.
Article
Computer Science, Artificial Intelligence
Leila Arras et al.
Summary: The rise of deep learning has increased the need to explain model decisions beyond prediction performances, leading to the development of XAI methods. Lack of objective quality measures for explanations has raised doubts on the trustworthiness of XAI methods. A new framework based on the CLEVR task was proposed in this study to evaluate ten different explanation methods, providing new insights on their quality and properties.
INFORMATION FUSION
(2022)
Article
Multidisciplinary Sciences
Avlant Nilsson et al.
Summary: The article introduces a method using artificial neural networks to simulate cellular signaling, predicting cellular responses to different signals and transcription factor activities. By processing experimental and simulated data from macrophages, researchers demonstrated the feasibility of this method in predicting intracellular signaling processes.
NATURE COMMUNICATIONS
(2022)
Article
Biochemical Research Methods
Gourab Ghosh Roy et al.
Summary: In this study, a visible neural network architecture based on biological pathway knowledge is proposed to improve the prediction of cancer survival risk. The Mutated Pathway Visible Neural Network (MPVNN) model is designed by randomly replacing known pathway edges using gene mutation data. Using the PI3K-Akt pathway as a case study, the MPVNN model demonstrates overall improved cancer-specific survival risk prediction compared to other similar-sized neural network models and standard survival analysis methods. The trained MPVNN architecture interpretation is reliable in identifying a small set of key genes within the PI3K-Akt pathway that is important in risk prediction for specific cancer types.
Proceedings Paper
Computer Science, Artificial Intelligence
Andreas Holzinger et al.
Summary: This article provides a brief overview of selected methods in the field of Explainable Artificial Intelligence (xAI), aiming to give beginners a quick summary of the current state of the art.
XXAI - BEYOND EXPLAINABLE AI: International Workshop, Held in Conjunction with ICML 2020, July 18, 2020, Vienna, Austria, Revised and Extended Papers
(2022)
Article
Computer Science, Artificial Intelligence
Yun Hao et al.
Summary: DTox is an interpretation framework for identifying toxicity pathways, predicting compound response to toxicity assays, improving interpretability, and discovering different mechanisms of toxicity.
Article
Biology
Arno van Hilten et al.
Summary: GenNet is an open-source deep learning framework that utilizes biologically knowledge to construct interpretable and memory-efficient neural network architectures, enabling the prediction of phenotypes and discovery of novel genes. It provides researchers with new insights into the genetic architecture of complex traits and diseases.
COMMUNICATIONS BIOLOGY
(2021)
Review
Physics, Multidisciplinary
Pantelis Linardatos et al.
Summary: Recent advances in artificial intelligence have led to widespread industrial adoption, with machine learning systems demonstrating superhuman performance. However, the complexity of these systems has made them difficult to explain, hindering their application in sensitive domains. Therefore, there is a renewed interest in the field of explainable artificial intelligence.
Article
Multidisciplinary Sciences
Haitham A. Elmarakeby et al.
Summary: The development of a biologically informed deep learning model P-NET enables stratification of prostate cancer patients by treatment-resistance state and evaluation of molecular drivers of treatment resistance for therapeutic targeting with complete model interpretability. This approach shows superior performance in cancer state prediction using molecular data compared to other modeling approaches.
Article
Multidisciplinary Sciences
John Jumper et al.
Summary: Proteins are essential for life, and accurate prediction of their structures is a crucial research problem. Current experimental methods are time-consuming, highlighting the need for accurate computational approaches to address the gap in structural coverage. Despite recent progress, existing methods fall short of atomic accuracy in protein structure prediction.
Article
Genetics & Heredity
Xiaoqing Huang et al.
Summary: This study developed a sparse learning framework called ParsVNN to build cancer-specific VNN models for predicting drug response in multiple cancer types. The results showed that the parsimony VNNs constructed by ParsVNN outperformed other methods in terms of prediction performance and identification of cancer driver genes. Additionally, the pathways selected by ParsVNN have great potential to predict clinical outcomes and recommend synergistic drug combinations.
NAR GENOMICS AND BIOINFORMATICS
(2021)
Article
Oncology
Brent M. Kuenzi et al.
Article
Biotechnology & Applied Microbiology
Nikolaus Fortelny et al.
Article
Multidisciplinary Sciences
Wassim Abida et al.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2019)
Article
Multidisciplinary Sciences
W. James Murdoch et al.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2019)
Article
Biochemistry & Molecular Biology
David A. Quigley et al.
Article
Biochemical Research Methods
Jianzhu Ma et al.
Article
Biochemical Research Methods
Jie Hao et al.
BMC BIOINFORMATICS
(2018)
Article
Multidisciplinary Sciences
Daifeng Wang et al.
Article
Multidisciplinary Sciences
Andre Esteva et al.
Article
Biochemistry & Molecular Biology
Ahmet Zehir et al.
Article
Medicine, Research & Experimental
Charles Dai et al.
COLD SPRING HARBOR PERSPECTIVES IN MEDICINE
(2017)
Article
Biochemistry & Molecular Biology
Jie Tan et al.
Review
Biochemistry & Molecular Biology
Christof Angermueller et al.
MOLECULAR SYSTEMS BIOLOGY
(2016)
Review
Multidisciplinary Sciences
Yann LeCun et al.
Article
Biochemistry & Molecular Biology
Nikolaus Fortelny et al.
Article
Biochemical Research Methods
Carito Guziolowski et al.
Article
Multidisciplinary Sciences
Jesse Gillis et al.
Article
Biochemistry & Molecular Biology
Julio Saez-Rodriguez et al.
MOLECULAR SYSTEMS BIOLOGY
(2009)
Review
Genetics & Heredity
AL Barabási et al.
NATURE REVIEWS GENETICS
(2004)
