metabolomics
ExoW3-NLDM
ExoW3-NLDM
54 LCMS runs (mzML). Every file is self-describing: the complete
experimental record is encoded directly into the mass
spectra with SpectraCodec. Decode any single file to
recover it. How the encoding and decoding works is documented in the
SpectraCodec repository.
Authenticity
Every file is cryptographically signed. The embedded message carries a
provenance block with two Ed25519 signatures: payload_signature covers the
embedded experimental record… See the full description on the dataset page: https://huggingface.co/datasets/lbnl-metabolomics/ExoW3-NLDM.edge-ML-metabolomics-graph
edge_ML metabolomics co-response graph
An undirected graph of 18,494 nodes and 2,709,209 edges built from pairwise
metabolite co-response statistics across 83 MetaboLights
studies, together with the node properties, the PyTorch Geometric graph object, and the
full pipeline that produces them.
A node is one differential comparison within one study assay (MTBLS1405_0002_00003332
= study MTBLS1405, assay 002, feature 00003332). An edge carries the association
between two… See the full description on the dataset page: https://huggingface.co/datasets/kozo2/edge-ML-metabolomics-graph.20210915_JGI-AK_MK_506588_SoilWaterRep_final_QE-HF_C18_USDAY63680
Soil water repellency (SWR) (i.e. soil hydrophobicity or decreased soil wettability) is a major cause of global soil degradation and a key agricultural concern. This metabolomics data will support the larger effort measuring soil water repellency and soil aggregate formation caused by microbial community composition through a combination of the standard drop penetration test, transmission electron microscopy characterization and physico-chemical analyses of soil aggregates at 6 timepoints.… See the full description on the dataset page: https://huggingface.co/datasets/lbnl-metabolomics/20210915_JGI-AK_MK_506588_SoilWaterRep_final_QE-HF_C18_USDAY63680.metabolomics_ptb_kd_resultsmetabolomics-edges-expected-ge5
Cross-study metabolomics co-response edges (expected frequency ≥ 5)
620,265 edges over 34,378 nodes, drawn from pairwise metabolite co-response
statistics across MetaboLights and
Metabolomics Workbench studies, together with
the node properties, two PyTorch Geometric graphs, and the full pipeline that produces
them.
A node is one differential comparison within one study assay — MTBLS1285_0001_00000028
is study MTBLS1285, assay 0001, feature 00000028; ST002832_AN004625_00002191… See the full description on the dataset page: https://huggingface.co/datasets/kozo2/metabolomics-edges-expected-ge5.