Artificial intelligence in the life sciences最新文献

Single cell RNA sequencing (scRNA-seq) technologies provide researchers with an unprecedented opportunity to exploit cell heterogeneity. For example, the sequenced cells belong to various cell lineages, which may have different cell fates in stem and progenitor cells. Those cells may differentiate into various mature cell types in a cell differentiation process. To trace the behavior of cell differentiation, researchers reconstruct cell lineages and predict cell fates by ordering cells chronologically into a trajectory with a pseudo-time. However, in scRNA-seq experiments, there are no cell-to-cell correspondences along with the time to reconstruct the cell lineages, which creates a significant challenge for cell lineage tracing and cell fate prediction. Therefore, methods that can accurately reconstruct the dynamic cell lineages and predict cell fates are highly desirable.

In this article, we develop an innovative machine-learning framework called Cell Smoothing Transformation (CellST) to elucidate the dynamic cell fate paths and construct gene networks in cell differentiation processes. Unlike the existing methods that construct one single bulk cell trajectory, CellST builds cell trajectories and tracks behaviors for each individual cell. Additionally, CellST can predict cell fates even for less frequent cell types. Based on the individual cell fate trajectories, CellST can further construct dynamic gene networks to model gene-gene relationships along the cell differentiation process and discover critical genes that potentially regulate cells into various mature cell types.

Natural products are attractive for drug discovery applications because of their distinctive chemical structures, such as an overall large fraction of sp³ carbon atoms, chiral centers (both features associated with structural complexity), large chemical scaffolds, and diversity of functional groups. Furthermore, natural products are used in de novo design and have inspired the development of pseudo-natural products using generative models. Public databases such as the Collection of Open NatUral ProdUcTs and the Universal Natural Product database (UNPD) are rich sources of structures to be used in generative models and other applications. In this work, we report the selection and characterization of the most diverse compounds of natural products from the UNPD using the MaxMin algorithm. The subsets generated with 14,994, 7,497, and 4,998 compounds are publicly available at https://github.com/DIFACQUIM/Natural-products-subsets-generation. We anticipate that the subsets will be particularly useful in building generative models based on natural products by research groups, particularly those with limited access to extensive supercomputer resources.

Ligand-based virtual screening plays an important role for cases in which protein structures are not available. Among ligand-based methods, accurate and fast prediction of protein-ligand binding affinity is crucial for reducing computational cost and exploring the chemical search space efficiently. Here we proposed a CNN-based method, termed as L3D-PLS for building the quantitative structure-activity relationships without target structures. In L3D-PLS, a CNN module was designed for extracting the key interaction features from the grids around aligned ligands, and a partial least square (PLS) model fits the binding affinity with the extracted features of the pre-trained CNN module. In 30 publicly available pre-aligned molecular datasets, L3D-PLS outperformed the traditional CoMFA method. This results highlight that L3D-PLS can be useful for lead optimization based on small datasets which is often true in drug discovery compaign.

The mechanism of action (MoA) of a compound describes the biological interaction through which it produces a pharmacological effect. Multiple data sources can be used for the purpose of predicting MoA, including compound structural information, and various assays, such as those based on cell morphology, transcriptomics and metabolomics. In the present study we explored the benefits and potential additive/synergistic effects of combining structural information, in the form of Morgan fingerprints, and morphological information, in the form of five-channel Cell Painting image data. For a set of 10 well represented MoA classes, we compared the performance of deep learning models trained on the two datasets separately versus a model trained on both datasets simultaneously. On a held-out test set we obtained a macro-averaged F1 score of 0.58 when training on only the structural data, 0.81 when training on only the image data, and 0.92 when training on both together. Thus indicating clear additive/synergistic effects and highlighting the benefit of integrating multiple data sources for MoA prediction.

One of the common strategies to identify novel chemical matter in drug discovery consists in performing a High Throughput Screening (HTS). However, the large amount of data generated at the dose-response (DR) step of an HTS campaign requires a careful analysis to detect artifacts and correct erroneous datapoints before validating the experiments. This step which requires to review each DR experiment can be time consuming and prone to human errors or inconsistencies. AI4DR is a system that has been developed for the classification of DR curves based on a Convolutional Neural Network (CNN) acting on normalized images of the DR curves. AI4DR allows the annotation in minutes of thousands of curves among 14 categories to help the High Throughput Screening biologists in their analyses. Several categories are associated with active and inactive compounds, other categories correspond to features of interest such as the presence of noise, a weaker effect at high doses, or a suspiciously weak or strong slope at the inflexion point of the DR curves of actives. The classifier has been trained on an algorithmically generated dataset curated and refined by experts, tested using real screening campaigns and improved using thousands of annotations by experts. The solution is deployed using a MLFlow model server interfaced with the Genedata Screener data analysis software used by the end users. AI4DR improves the consistency, the robustness, and the speed of HTS data analysis as well as reducing the human effort to identify faster new medicines for patients.

Recent advances in the field of artificial intelligence, specifically regarding deep learning methods, have invigorated research into novel ways for the exploration of chemical space. Compared to more traditional methods that rely on chemical fragments and combinatorial recombination deep generative models generate molecules in a non-transparent way that defies easy rationalization. However, this opaque nature also promises to explore uncharted chemical space in novel ways that do not rely on structural similarity directly. These aspects and the complexity of training such models makes model assessment regarding novelty, uniqueness, and distribution of generated molecules a central aspect. This perspective gives an overview of current methodologies for chemical space exploration with an emphasis on deep neural network approaches. Key aspects of generative models include choice of molecular representation, the targeted chemical space, and the methodology for assessing and validating chemical space coverage.

Artificial intelligence is becoming an important diagnostic and prognostic tool in recent years, as machine learning algorithms have been shown to improve clinical decision-making. These algorithms will have some of their most important applications in developing regions with restricted data collection, but their performance under this condition is still widely unknown. We analyzed longitudinal data from São Paulo, Brazil, to develop machine learning algorithms to predict the risk of death in patients with cancer. We tested different algorithms using nine separate model structures. Considering the area under the ROC curve (AUC-ROC), we obtained values of 0.946 for the general model, 0.945 for the model with the five main cancers, 0.899 for bronchial and lung cancer, 0.947 for breast cancer, 0.866 for stomach cancer, 0.872 for colon cancer, 0.923 for rectum cancer, 0.955 for prostate cancer, and 0.917 for uterine cervix cancer. Our results indicate the potential of building models for predicting mortality risk in cancer patients in developing regions using only routinely-collected data.

One of the main computational tools for structure-based drug discovery is molecular docking. Due to the natural representation of molecules as graphs (a set of nodes/atoms connected through edges/bonds), Deep Graph Learning has been successfully applied for multiple tasks on this area. This work presents an overview of Deep Graph Learning methods developed within this research field, as well as opportunities for future development.

Ontologies are used to support access to a multitude of databases that cover domains relevant information. Heterogeneity and different semantics can be accessed by using structured texts and descriptions in a hierarchical concept definition. We are interested in Life Sciences (LS) related ontologies including components taken from molecular biology, bioinformatics, physics, chemistry, medicine and other related areas. An Ontology comprises: (i) term connections, (ii) the identification of core concepts, (iii) data management, (iv) knowledge classification and integration to collect key information. An ontology may be very useful in navigating through LS terms. This paper explores some available biomedical ontologies and frameworks. It describes the most common ontology development environments (ODE): Protégé, Topbraid Composer, Ontostudio, Fluent Editor, VocBench, Swoop and Obo-edit, to create ontologies from textual scientific resources for LS plans. It also compares ontology methodologies in terms of Usability, Scalability, Stability, Integration, Documentation and Originality.