Cell-Type-Resolved Ligand-Receptor Connectome (CLRC)

This dataset contains the cell-type-resolved ligand–receptor connectome (CLRC) described in Pak, Hong et al., Cell-Type-Resolved Ligand-Receptor Connectome of the Human Brain and Its Disruption in Alzheimer's Disease. The CLRC maps 1,014 ligand–receptor pairs across 2,133 region–cell-type nodes derived from the Allen Brain Cell (ABC) Atlas, yielding directed communication matrices that characterize cell-type-specific signalling across the human brain. Contents clrc_gpu.h5 (119.3 GB, HDF5) — ligand–receptor communication scores computed with NeuronChat. Conventions M[i, j] = signalling from node i (ligand/sender) to node j (receptor/receiver). Directed and asymmetric. Nodes: 2,133 observed region–cell-type combinations (109 ABC dissections × 31 superclusters, e.g. MTG-Oligodendrocyte). Includes intra- and inter-regional edges. Values: NeuronChat communication strengths (trimean-summarized log₂ expression → geometric mean → Hill function). 0 = tested but not significant (permutation P ≥ 0.05). NaN = ligand or receptor not detected. These mean different things — don't collapse them. Usage Read slices rather than loading datasets into memory in full: import h5py with h5py.File("clrc_gpu.h5", "r") as f: print(list(f.keys())) mat = f["[dataset name]"][0] Derivation Communication scores were computed with NeuronChat from single-nucleus transcriptomic data in the Allen Brain Cell Atlas. Analysis code: github.com/neuropm-lab/CLRC Attribution Derived from the Allen Brain Cell Atlas, © Allen Institute for Brain Science, available under CC-BY-4.0. Please cite the original atlas alongside this dataset.