Large scale similarity search across digital reconstructions of neural morphology

Accelerating the continuous community sharing of digital neuromorphology data

The tree-like morphology of neurons and glia is a key cellular determinant of circuit connectivity and metabolic function in the nervous system of essentially all animals. To elucidate the contribution of specific cell types to both physiological and pathological brain states, it is important to access detailed neuroanatomy data for quantitative analysis and computational modeling. NeuroMorpho.Org is the largest online collection of freely available digital neural reconstructions and related metadata and is continuously updated with new uploads. Earlier in the project, we released multiple datasets together yearly, but this process caused an average delay of several months in making the data public. Moreover, in the past 5 years, >80% of invited authors agreed to share their data with the community via NeuroMorpho.Org, up from <20% in the first 5 years of the project. In the same period, the average number of reconstructions per publication increased 600%, creating the need for automatic processing to release more reconstructions in less time. The progressive automation of our pipeline enabled the transition to agile releases of individual datasets as soon as they are ready. The overall time from data identification to public sharing decreased by 63.7%; 78% of the datasets are now released in less than 3 months with an average workflow duration below 40 days. Furthermore, the mean processing time per reconstruction dropped from 3 h to 2 min. With these continuous improvements, NeuroMorpho.Org strives to forge a positive culture of open data. Most importantly, the new, original research enabled through reuse of datasets across the world has a multiplicative effect on science discovery, benefiting both authors and users.

FaissMolLib: An efficient and easy deployable tool for ligand-based virtual screening

Virtual screening-based molecular similarity and fingerprint are crucial in drug design, target prediction, and ADMET prediction, aiding in identifying potential hits and optimizing lead compounds. However, challenges such as lack of comprehensive open-source molecular fingerprint databases and efficient search methods for virtual screening are prevalent. To address these issues, we introduce FaissMolLib, an open-source virtual screening tool that integrates 2.8 million compounds from ChEMBL and ZINC databases. Notably, FaissMolLib employs the highly efficient Faiss search algorithm, outperforming the Tanimoto algorithm in identifying similar molecules with its tighter clustering in scatter plots and lower mean, standard deviation, and variance in key molecular properties. This feature enables FaissMolLib to screen 2.8 million compounds in just 0.05 seconds, offering researchers an efficient, easily deployable solution for virtual screening on laptops and building unique compound databases. This significant advancement holds great potential for accelerating drug discovery efforts and enhancing chemical data analysis. FaissMolLib is freely available at http://liuhaihan.gnway.cc:80. The code and dataset of FaissMolLib are freely available at https://github.com/Superhaihan/FiassMolLib.

Automatic identification of scientific publications describing digital reconstructions of neural morphology

The increasing number of peer-reviewed publications constitutes a challenge for biocuration. For example, NeuroMorpho.Org, a sharing platform for digital reconstructions of neural morphology, must evaluate more than 6000 potentially relevant articles per year to identify data of interest. Here, we describe a tool that uses natural language processing and deep learning to assess the likelihood of a publication to be relevant for the project. The tool automatically identifies articles describing digitally reconstructed neural morphologies with high accuracy. Its processing rate of 900 publications per hour is not only amply sufficient to autonomously track new research, but also allowed the successful evaluation of older publications backlogged due to limited human resources. The number of bio-entities found since launching the tool almost doubled while greatly reducing manual labor. The classification tool is open source, configurable, and simple to use, making it extensible to other biocuration projects.

BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. We observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

Cell morphologies in the nervous system: Glia steal the limelight

Neurons and glia have distinct yet interactive functions but are both characterized by branching morphology. Dendritic trees have been digitally traced for over 40 years in many animal species, anatomical regions, and neuron types. Recently, long-range axons also are being reconstructed throughout the brain of many organisms from invertebrates to primates. In contrast, less attention has been paid until lately to glial morphology. Thus, although glia and neurons are similarly abundant in the nervous systems of humans and most animal models, glia have traditionally been much less represented than neurons in morphological reconstruction repositories such as NeuroMorpho.Org. This is rapidly changing with the advent of high-throughput glia tracing. NeuroMorpho.Org introduced glial cells in 2017 and today they constitute nearly a third of the database content. It took NeuroMorpho.Org 10 years to collect the first 40,000 neurons and now that amount of data can be produced in a single publication. This not only demonstrates the spectacular technological progress in data production, but also demands a corresponding advancement in informatics processing. At the same time, these publicly available data also open new opportunities for quantitative analysis and computational modeling to identify universal or cell-type-specific design principles in the cellular architecture of nervous systems. As a first application, we demonstrated that supervised machine learning of tree geometry classifies neurons and glia with practically perfect accuracy. Furthermore, we discovered a new morphometric biomarker capable of robustly separating these cell classes across multiple species, brain regions, and experimental preparations, with only sparse sampling of branch measurements.

Neuron tracing from light microscopy images: automation, deep learning and bench testing

MOTIVATION

Large-scale neuronal morphologies are essential to neuronal typing, connectivity characterization and brain modeling. It is widely accepted that automation is critical to the production of neuronal morphology. Despite previous survey papers about neuron tracing from light microscopy data in the last decade, thanks to the rapid development of the field, there is a need to update recent progress in a review focusing on new methods and remarkable applications.

RESULTS

This review outlines neuron tracing in various scenarios with the goal to help the community understand and navigate tools and resources. We describe the status, examples and accessibility of automatic neuron tracing. We survey recent advances of the increasingly popular deep-learning enhanced methods. We highlight the semi-automatic methods for single neuron tracing of mammalian whole brains as well as the resulting datasets, each containing thousands of full neuron morphologies. Finally, we exemplify the commonly used datasets and metrics for neuron tracing bench testing.

Assisted neuroscience knowledge extraction via machine learning applied to neural reconstruction metadata on NeuroMorpho.Org

The amount of unstructured text produced daily in scholarly journals is enormous. Systematically identifying, sorting, and structuring information from such a volume of data is increasingly challenging for researchers even in delimited domains. Named entity recognition is a fundamental natural language processing tool that can be trained to annotate, structure, and extract information from scientific articles. Here, we harness state-of-the-art machine learning techniques and develop a smart neuroscience metadata suggestion system accessible by both humans through a user-friendly graphical interface and machines via Application Programming Interface. We demonstrate a practical application to the public repository of neural reconstructions, NeuroMorpho.Org, thus expanding the existing web-based metadata management system currently in use. Quantitative analysis indicates that the suggestion system reduces personnel labor by at least 50%. Moreover, our results show that larger training datasets with the same software architecture are unlikely to further improve performance without ad-hoc heuristics due to intrinsic ambiguities in neuroscience nomenclature. All components of this project are released open source for community enhancement and extensions to additional applications.