Functional neuroanatomical maps provide a mesoscale reference framework for studies from molecular to systems neuroscience and psychiatry. The underlying structure-function relationships are typically derived from functional manipulations or imaging approaches. Although highly informative, these are experimentally costly. The increasing amount of publicly available brain and genetic data offers a rich source that could be mined to address this problem computationally. Here, we developed an algorithm that fuses gene expression and connectivity data with functional genetic meta data and exploits cumulative effects to derive neuroanatomical maps related to multi-genic functions. We validated the approach by using public available mouse and human data. The generated neuroanatomical maps recapture known functional anatomical annotations from literature and functional MRI data. When applied to multi-genic meta data from mouse quantitative trait loci (QTL) studies and human neuropsychiatric databases, this method predicted known functional maps underlying behavioral or psychiatric traits. Taken together, genetically weighted connectivity analysis (GWCA) allows for high throughput functional exploration of brain anatomy in silico. It maps functional genetic associations onto brain circuitry for refining functional neuroanatomy, or identifying trait-associated brain circuitry, from genetic data.

Motivation The emergence of single-cell RNA sequencing (scRNA-seq) has led to an explosion in novel methods to study biological variation among individual cells, and to classify cells into functional and biologically meaningful categories. Results Here, we present a new cell type projection tool, HieRFIT (Hierarchical Random Forest for Information Transfer), based on hierarchical random forests. HieRFIT uses a priori information about cell type relationships to improve classification accuracy, taking as input a hierarchical tree structure representing the class relationships, along with the reference data. We use an ensemble approach combining multiple random forest models, organized in a hierarchical decision tree structure. We show that our hierarchical classification approach improves accuracy and reduces incorrect predictions especially for inter-dataset tasks which reflect real life applications. We use a scoring scheme that adjusts probability distributions for candidate class labels and resolves uncertainties while avoiding the assignment of cells to incorrect types by labeling cells at internal nodes of the hierarchy when necessary. Availability HieRFIT is implemented as an R package, and it is available at (https://github.com/yasinkaymaz/HieRFIT/releases/tag/v1.0.0). t Supplementary information Supplementary data are available at Bioinformatics online.

No abstract

Recent advances in neuro-imaging allowed big brain-initiatives and consortia to create vast resources of brain data that can be mined by researchers for their individual projects. Exploring the relationship between genes, brain circuitry, and behavior is one of key elements of neuroscience research. This requires fusion of spatial connectivity data at varying scales, such as whole brain correlated gene expression, structural and functional connectivity. With ever-increasing resolution, those exceed the past state-of-the art in several orders of magnitude in size and complexity. Current analytical workflows in neuroscience involve time-consuming manual aggregation of the data and only sparsely incorporate spatial context to operate continuously on multiple scales. Incorporating techniques for handling big connectivity data is therefore a necessity.We propose a data structure to explore heterogeneous neurobiological connectivity data for integrated visual analytics workflows. Aggregation Queries, i.e. the aggregated connectivity from, to or between brain areas allow experts the comparison of multimodal networks residing at different scales, or levels of hierarchically organized anatomical atlases. Executed on-demand on volumetric gene expression and connectivity data, they enable an interactive dissection of networks, with billions of edges, in real-time, and based on their spatial context. The data structure is optimized to be accessed directly from the hard disk, since connectivity of large-scale networks typically exceed the memory size of current consumer level PCs. This allows experts to embed and explore their own experimental data in the framework of public data resources without large-scale infrastructure.Our novel data structure outperforms state-of-the-art graph engines in retrieving connectivity of local brain areas experimentally. We demonstrate the application of our approach for neuroscience by analyzing fear-related functional neuroanatomy in mice. Further, we show its versatility by comparing multimodal brain networks linked to autism. Importantly, we achieve cross-species congruence in retrieving human psychiatric traits networks, which facilitates selection of neural substrates to be further studied in mouse models.

No abstract