Artificial intelligence in the life sciences最新文献

Artificial intelligence (AI) based on deep learning (DL) has sparked tremendous global interest in recent years. Deep Learning has been widely adopted in speech and image recognition, natural language processing which has an impact on healthcare. In the recent decade, the application of DL has exponentially grown in the field of Ophthalmology. The fundoscopy, slit lamp photography, optical coherence tomography (OCT), and magnetic resonance imaging (MRI) were employed for clinical examination of various ocular conditions. These data served as a perfect platform for the development of DL models in Ophthalmology. Currently, the application of DL in ocular disorders is majorly studied in Diabetic retinopathy (DR), age-related macular degeneration (AMD), macular oedema, retinopathy of prematurity (ROP), glaucoma, and cataract. In Ophthalmology, DL models are gradually expanding their scope in optic neuropathies. Glaucoma and optic neuritis are optic nerve disorders, where DL models are currently studied for clinical applications. For further expansion of DL application in inherited optic neuropathies, we discussed the recent observational studies revealing the pathophysiological changes at the optic nerve in Leber's hereditary optic neuropathy (LHON). LHON is an inherited optic neuropathy leading to bilateral loss of vision in early age groups. Hence for early management, further footsteps in the application of DL in LHON will benefit both ophthalmologists and patients. In this review, we discuss the recent advancements of AI in the Ophthalmology and prospective of applying DL models in LHON for clinical precision and timely diagnosis.

Despite the burgeoning development of artificial intelligence (AI) applied in the medical field, there have been little bibliometric and collaboration network researches on the patents related to this inter-disciplinary research domain. Patentometric and Social Network Analysis (SNA) are used to conduct the characterizations of patent applications and cooperative networks, mapping a holistic landscape related to the AI-medical field. Derwent Innovation Index database (DII) is adopted as the patent data source. The results indicate that the quantity of AI-medical-related patent applications has been increasing explosively since 2011. The United States of America (US) is both the foremost country developing related technologies and the primary target of patent filing by non-residents. The hotspot of the current research include medical image recognition, computer-aided diagnosis, disease monitoring, disease prediction, bioinformatics, and drug development, etc. Low density of the assignees cooperation network implies the slight patent collaboration. Companies and academic institutions are the friskiest innovation subjects in the AI-medical field. The geographical proximity has a positive influence on the patent collaboration because co-owned patents are concentrated on the institutes in the same nation. Domestic collaboration is the major collaborative pattern. The spatial agglomeration of trans-regional patent cooperation is fairly sparse, which requires a further escalation in knowledge circulation. It has practical significance to understand the developing situation and patent cooperation network in the AI-medical field, providing a reference for future strategy planning, development, and technological marketization.

The objective of this article is to review the application of informatics-driven approaches in the pharmacovigilance field with focus on drug-drug interaction (DDI) safety signal discovery using various data sources. Signal can be a new safety information or new aspect to already known adverse drug reaction which is possibly causally related to a medication/medications that warrants further investigation to accept or refute. Signals can be detected from different data sources such as spontaneous reporting system, scientific literature, biomedical databases and electronic health records. This review is substantiated based on the fact that DDIs are contributing to a public health problem represented in 6-30% adverse drug event occurrences. In this article, we review informatics-driven approaches applied by authors focusing on DDI signal detection using different data sources. The aim of this article is not to laboriously survey all PV literature. As an alternative, we discussed informatics-driven methods used to discover DDI signals and various data sources reinforced with instances of studies from PV literature. The adoption of informatics-driven approaches can complement and optimize the practice of safety signal detection. However, further researches should be carried out to evaluate the efficiency of those approaches and to address the limitations of external validation, implementation and adoption in real clinical environments and by the regulatory bodies.

The lack of bioinformatics tools to quickly assess protein conformational and topological features motivated to create an integrative and user-friendly R package. Moreover, the Fiscore package implements a pipeline for Gaussian mixture modelling making such machine learning methods readily accessible to non-experts. This is especially important since probabilistic machine learning techniques can help with a better interpretation of complex biological phenomena when it is necessary to elucidate various structural features that might play a role in protein function. Thus, Fiscore builds on the mathematical formulation of protein physicochemical properties that can aid in drug discovery, target evaluation, or relational database building. In addition, the package provides interactive environments to explore various features of interest. Finally, one of the goals of this package was to engage structural bioinformaticians and develop more robust and free R tools that could help researchers not necessarily specialising in this field. Package Fiscore (v.0.1.3) is distributed free of charge via CRAN and Github.

Deep learning is very promising in solving problems in omics research, such as genomics, epigenomics, proteomics, and metabolics. The design of neural network architecture is very important in modeling omics data against different scientific problems. Residual fully-connected neural network (RFCN) was proposed to provide better neural network architectures for modeling omics data. The next challenge for omics research is how to integrate information from different omics data using deep learning, so that information from different molecular system levels could be combined to predict the target. In this paper, we present a novel multi-omics integration approach named AutoOmics that could efficiently integrate information from different omics data and achieve better accuracy than previous approaches. We evaluated our method on four different tasks: drug repositioning, target gene prediction, breast cancer subtyping and cancer type prediction, and all the four tasks achieved state of art performances.

Bile salt export pump (BSEP) is a transporter that moves bile salts from hepatocytes into bile canaliculi. BSEP inhibition can result in the toxic accumulation of bile salts in the liver, which has been identified as a risk factor of drug-induced liver injury (DILI). Since DILI is a frequent cause of drug withdrawals from the market or failings in drug development, in vitro BSEP activity is measured with the [³H]taurocholate uptake assay and a half-maximal inhibitory concentration (IC₅₀) higher than 30 µM is advised. Herein, a machine learning classification model was developed to accurately detect BSEP inhibitors and help in the prioritization of in vitro testing. Regression models for the numerical prediction of IC₅₀ values were also generated. Classification and regression models for BSEP inhibition have been evaluated on realistic settings, which is critical prior to ML-based decision making in drug discovery programs. This work illustrates how predictive safety can help in early toxicity risk assessment and compound prioritization by leveraging Novartis historical experimental data.

Compounds interfering with high-throughput screening (HTS) assay technologies (also known as “badly behaving compounds”, “bad actors”, “nuisance compounds” or “PAINS”) pose a major challenge to early-stage drug discovery. Many of these problematic compounds are “frequent hitters”, and we have recently published a set of machine learning models (“Hit Dexter 2.0”) for flagging such compounds.

Here we present a new generation of machine learning models which are derived from a large, manually curated and annotated data set. For the first time, these models cover, in addition to target-based assays, also cell-based assays. Our experiments show that cell-based assays behave indeed differently from target-based assays, with respect to hit rates and frequent hitters, and that dedicated models are required to produce meaningful predictions. In addition to these extensions and refinements, we explored a variety of additional setups for modeling, including the combination of four machine learning classifiers (i.e. k-nearest neighbors (KNN), extra trees, random forest and multilayer perceptron) with four sets of descriptors (Morgan2 fingerprints, Morgan3 fingerprints, MACCS keys and 2D physicochemical property descriptors).

Testing on holdout data as well as data sets of “dark chemical matter” (i.e. compounds that have been extensively tested in biological assays but have never shown activity) and known bad actors show that the multilayer perceptron classifiers in combination with Morgan2 fingerprints outperform other setups in most cases. The best multilayer perceptron classifiers obtained Matthews correlation coefficients of up to 0.648 on holdout data. These models are available via a free web service.