Tools and Analysis

NeMO - Broad Transcriptome Analysis

 

 

Explore the BICCN data with a tutorial.

For BICCN data access in FireCloud,  a quick registration (< 1 minute) is required.

A main goal of the BCDC is bringing BICCN data to the community and to facilitate its use and analysis.  To help users get started we have prepared a sample analysis workflow whereby you can

Moving forward, we will automate and increase the integration of the NeMO and FireCloud/Single Cell Portal Environments. Currently, this demo focuses on running single cell transcriptomics data (specifically 10X v2).

 


 

 

Cell Type Tools

The BCDC is committed to making tools available to the community that are relevant for the analysis of cell types.   If you have a tool you think would be useful please write to info@biccn.org.   As the number of entries grow we will categorize them.

 

L-Measure is software tool designed to extract a wide variety of quantitative morphological measurements from neuromorphological reconstructions. Both local parameters (e.g. bifurcation angles) and global descriptors (e.g. total arbor length) can be extracted and combined in many useful analyses, including the popular distributions of surface area as a function of path distance from the soma. Users can specify the target of the analysis by structural domains (e.g. axons vs dendrites) or by morphological features (e.g. terminal branches) . The tool has built-in capability to search for neurons with specific morphological characteristics from a large collection or to compare two neuronal populations with parametric and non-parametric statistical tests. The user-friendly graphical user interface is written in JAVA and can run remotely through a web browser or locally on Linux, Windows, or Mac. The number-crunching engine is written in C++ and can be called from batch scripts for faster execution of large-scale computations.  

https://www.nature.com/articles/nprot.2008.51

G. Ascoli Lab ( http://krasnow1.gmu.edu/cn3/ascoli/), George Mason University  

 


 

 

Loom is an efficient file format for very large omics datasets, consisting of a main matrix, optional additional layers, a variable number of row and column annotations, and sparse graph objects. Loom files are portable, self-contained and ensure that metadata travels with the data. Under the hood, Loom files are HDF5 and can be opened from many programming languages, including Python, R, C, C++, Java, MATLAB, Mathematica, and Julia. The Loom file format is natively supported by popular scRNA-seq packages including Scanpy, Seurat, SCope and scVI.

 

https://loompy.org

S. Linnarson Lab (http://linnarssonlab.org/), Karolinska Institute 


 

 

MetaNeighbor quantifies the degree to which cell types replicate across datasets, and enables rapid identification of clusters with high similarity. MetaNeighbor first measures the replicability of neuronal identity, comparing results across eight technically and biologically diverse datasets to define best practices for more complex assessments.   By taking the correlations between all pairs of cells a network is built where every node is a cell and the edges represent how similar each cell is to each other cell.  This network can be extended to include data from multiple experiments (multiple datasets).  To assess cell-type identity across experiments neighbor voting is used for cross-validation, systematically hiding the labels from one dataset at a time for testing. Cells within the test set are predicted as similar to the cell types from other training sets using a neighbor-voting formalism. Whether these scores prioritize cells as the correct type within the dataset determines the performance, expressed as the AUROC. Comparative assessment of cells occurs only within a dataset, but is based only on training information from outside that dataset. 

 

https://www.bioconductor.org/packages/release/bioc/html/MetaNeighbor.html

J. Gillis Lab (http://gillislab.labsites.cshl.edu/), Cold Spring Harbor Laboratory


 

 

Biophysical models. The advanced cognitive capabilities of the human brain are often attributed to our recently evolved neocortex. However, it is not known whether the basic building blocks of human neocortex, the pyramidal neurons, possess unique biophysical properties that might impact on cortical computations. The Segev group has shown that layer 2/3 pyramidal neurons from human temporal cortex (HL2/3 PCs) have a specific membrane capacitance (Cm) of ~0.5 µF/cm2, half of the commonly accepted “universal” value (~1 µF/cm2) for biological membranes. This finding was predicted by fitting in vitro voltage transients to theoretical transients then validated by direct measurement of Cm in nucleated patch experiments. This is the first demonstration that human cortical neurons have distinctive membrane properties, suggesting important implications for signal processing in human neocortex. They also have developed detailed models of pyramidal cells from human neocortex, including models on their excitatory synapses, dendritic spines, dendritic NMDA- and somatic/axonal Na+ spikes that provided new insights into signal processing and computational capabilities of these principal cells.     All models and code are available at the links below. 

https://senselab.med.yale.edu/ModelDB/showmodel.cshtml?model=195667#tabs-1

https://senselab.med.yale.edu/ModelDB/showmodel.cshtml?model=238347#tabs-1

 

I. Segev Lab (https://elsc.huji.ac.il/segev/home), Hebrew University