head_id stringlengths 6 51 | head_type stringclasses 5
values | relation stringclasses 25
values | tail_id stringlengths 7 79 | tail_type stringclasses 5
values | confidence stringclasses 97
values | species_source stringclasses 10
values | source stringclasses 182
values | evidence stringlengths 16 14k | evidence_type stringclasses 10
values |
|---|---|---|---|---|---|---|---|---|---|
D000006 | disease | is_a | D015746 | disease | mesh_tree | human | MeSH | tree=C23.888.592.612.054.200>C23.888.592.612.054|tree=C23.888.821.030.249>C23.888.821.030 | ontology |
D000007 | disease | is_a | D014947 | disease | mesh_tree | human | MeSH | tree=C26.017>C26 | ontology |
D000008 | disease | is_a | D009371 | disease | mesh_tree | human | MeSH | tree=C04.588.033>C04.588 | ontology |
D000012 | disease | is_a | D006995 | disease | mesh_tree | human | MeSH | tree=C16.320.565.398.500.440.500>C16.320.565.398.500.440|tree=C18.452.584.500.875.440.500>C18.452.584.500.875.440|tree=C18.452.584.563.500.440.500>C18.452.584.563.500.440|tree=C18.452.648.398.500.440.500>C18.452.648.398.500.440 | ontology |
D000013 | disease | is_a | D009358 | disease | mesh_tree | human | MeSH | tree=C16.131>C16 | ontology |
D000014 | disease | is_a | D000013 | disease | mesh_tree | human | MeSH | tree=C16.131.042>C16.131 | ontology |
D000015 | disease | is_a | D000013 | disease | mesh_tree | human | MeSH | tree=C16.131.077>C16.131 | ontology |
D000016 | disease | is_a | D000013 | disease | mesh_tree | human | MeSH | tree=C16.131.080>C16.131 | ontology |
D000016 | disease | is_a | D011832 | disease | mesh_tree | human | MeSH | tree=C26.733.031>C26.733 | ontology |
D000022 | disease | is_a | D011248 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039>C12.050.703 | ontology |
D000026 | disease | is_a | D000022 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039.089>C12.050.703.039 | ontology |
D000027 | disease | is_a | D000022 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039.093>C12.050.703.039 | ontology |
D000030 | disease | is_a | D000022 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039.173>C12.050.703.039 | ontology |
D000031 | disease | is_a | D000022 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039.256>C12.050.703.039 | ontology |
D000031 | disease | is_a | D011251 | disease | mesh_tree | human | MeSH | tree=C01.674.173>C01.674|tree=C12.050.703.700.173>C12.050.703.700 | ontology |
D000033 | disease | is_a | D011248 | disease | mesh_tree | human | MeSH | tree=C12.050.703.090>C12.050.703 | ontology |
D000034 | disease | is_a | D000022 | disease | mesh_tree | human | MeSH | tree=C12.050.703.039.422>C12.050.703.039 | ontology |
D000034 | disease | is_a | D000820 | disease | mesh_tree | human | MeSH | tree=C22.021>C22 | ontology |
D000037 | disease | is_a | D007744 | disease | mesh_tree | human | MeSH | tree=C12.050.703.420.078>C12.050.703.420 | ontology |
D000037 | disease | is_a | D010922 | disease | mesh_tree | human | MeSH | tree=C12.050.703.590.132>C12.050.703.590 | ontology |
D000038 | disease | is_a | D013492 | disease | mesh_tree | human | MeSH | tree=C01.830.025>C01.830|tree=C23.550.470.756.100>C23.550.470.756 | ontology |
D000039 | disease | is_a | D000038 | disease | mesh_tree | human | MeSH | tree=C01.830.025.675>C01.830.025 | ontology |
D000039 | disease | is_a | D014069 | disease | mesh_tree | human | MeSH | tree=C01.748.561.750.500>C01.748.561.750|tree=C07.550.781.750.500>C07.550.781.750|tree=C08.730.561.750.500>C08.730.561.750|tree=C09.775.649.750.500>C09.775.649.750 | ontology |
D000051 | disease | is_a | D010335 | disease | mesh_tree | human | MeSH | tree=C23.550.035>C23.550 | ontology |
D000051 | disease | is_a | D012872 | disease | mesh_tree | human | MeSH | tree=C17.800.865.070>C17.800.865 | ontology |
D000052 | disease | is_a | D008548 | disease | mesh_tree | human | MeSH | tree=C17.800.621.430.530.100>C17.800.621.430.530 | ontology |
D000067011 | disease | is_a | D044342 | disease | mesh_tree | human | MeSH | tree=C18.654.521.719>C18.654.521 | ontology |
D000067073 | disease | is_a | D040921 | disease | mesh_tree | human | MeSH | tree=F03.950.750.375>F03.950.750 | ontology |
D000067208 | disease | is_a | D005512 | disease | mesh_tree | human | MeSH | tree=C20.543.480.370.763>C20.543.480.370 | ontology |
D000067251 | disease | is_a | D012008 | disease | mesh_tree | human | MeSH | tree=C23.550.291.937.500>C23.550.291.937 | ontology |
D000067329 | disease | is_a | D009765 | disease | mesh_tree | human | MeSH | tree=C18.654.726.750.500.650>C18.654.726.750.500|tree=C23.888.144.699.500.250>C23.888.144.699.500 | ontology |
D000067390 | disease | is_a | D014947 | disease | mesh_tree | human | MeSH | tree=C26.212>C26 | ontology |
D000067398 | disease | is_a | D014947 | disease | mesh_tree | human | MeSH | tree=C26.946>C26 | ontology |
D000067404 | disease | is_a | D003147 | disease | mesh_tree | human | MeSH | tree=F03.625.374.250>F03.625.374 | ontology |
D000067454 | disease | is_a | D003147 | disease | mesh_tree | human | MeSH | tree=F03.625.374.125>F03.625.374 | ontology |
D000067490 | disease | is_a | D063487 | disease | mesh_tree | human | MeSH | tree=C25.775.342.500.750>C25.775.342.500|tree=F03.900.384.500.750>F03.900.384.500 | ontology |
D000067559 | disease | is_a | D007859 | disease | mesh_tree | human | MeSH | tree=C10.597.606.150.550.700>C10.597.606.150.550|tree=C23.888.592.604.150.550.700>C23.888.592.604.150.550|tree=F03.625.374.188.700>F03.625.374.188|tree=F03.625.562.700>F03.625.562 | ontology |
D000067562 | disease | is_a | D020969 | disease | mesh_tree | human | MeSH | tree=C23.550.291.883>C23.550.291 | ontology |
D000067836 | disease | is_a | D009771 | disease | mesh_tree | human | MeSH | tree=F03.080.600.250>F03.080.600 | ontology |
D000067877 | disease | is_a | D002659 | disease | mesh_tree | human | MeSH | tree=F03.625.164.113>F03.625.164 | ontology |
D000068079 | disease | is_a | D001523 | disease | mesh_tree | human | MeSH | tree=F03.608>F03 | ontology |
D000068099 | disease | is_a | D001523 | disease | mesh_tree | human | MeSH | tree=F03.950>F03 | ontology |
D000068105 | disease | is_a | D019964 | disease | mesh_tree | human | MeSH | tree=F03.600.150>F03.600 | ontology |
D000068116 | disease | is_a | D020018 | disease | mesh_tree | human | MeSH | tree=F03.835.550>F03.835 | ontology |
D000068376 | disease | is_a | D005222 | disease | mesh_tree | human | MeSH | tree=C23.888.369.500.500>C23.888.369.500 | ontology |
D000069076 | disease | is_a | D009104 | disease | mesh_tree | human | MeSH | tree=C26.640.500>C26.640 | ontology |
D000069076 | disease | is_a | D050723 | disease | mesh_tree | human | MeSH | tree=C26.404.280>C26.404 | ontology |
D000069279 | disease | is_a | D004827 | disease | mesh_tree | human | MeSH | tree=C10.228.140.490.125>C10.228.140.490 | ontology |
D000069281 | disease | is_a | D000072659 | disease | mesh_tree | human | MeSH | tree=C10.228.140.617.738.275.500>C10.228.140.617.738.275|tree=C19.700.419.500>C19.700.419 | ontology |
D000069281 | disease | is_a | D001327 | disease | mesh_tree | human | MeSH | tree=C20.111.273>C20.111 | ontology |
D000069282 | disease | is_a | D003607 | disease | mesh_tree | human | MeSH | tree=C11.496.221.500>C11.496.221 | ontology |
D000069290 | disease | is_a | D006547 | disease | mesh_tree | human | MeSH | tree=C23.300.707.945>C23.300.707 | ontology |
D000069290 | disease | is_a | D011183 | disease | mesh_tree | human | MeSH | tree=C23.550.767.500>C23.550.767 | ontology |
D000069293 | disease | is_a | D016403 | disease | mesh_tree | human | MeSH | tree=C04.557.386.480.150.585.500>C04.557.386.480.150.585|tree=C15.604.515.569.480.150.585.500>C15.604.515.569.480.150.585|tree=C20.683.515.761.480.150.585.500>C20.683.515.761.480.150.585 | ontology |
D000069295 | disease | is_a | D002277 | disease | mesh_tree | human | MeSH | tree=C04.557.470.200.588>C04.557.470.200 | ontology |
D000069316 | disease | is_a | D000152 | disease | mesh_tree | human | MeSH | tree=C17.800.030.150.500>C17.800.030.150|tree=C17.800.794.111.500>C17.800.794.111 | ontology |
D000069337 | disease | is_a | D014564 | disease | mesh_tree | human | MeSH | tree=C12.050.351.875.420>C12.050.351.875|tree=C12.200.706.445>C12.200.706|tree=C12.800.445>C12.800|tree=C16.131.939.445>C16.131.939 | ontology |
D000069451 | disease | is_a | D010335 | disease | mesh_tree | human | MeSH | tree=C23.550.543>C23.550 | ontology |
D000069544 | disease | is_a | D002494 | disease | mesh_tree | human | MeSH | tree=C01.207.399>C01.207|tree=C10.228.228.399>C10.228.228 | ontology |
D000069544 | disease | is_a | D004660 | disease | mesh_tree | human | MeSH | tree=C10.228.140.430.520>C10.228.140.430|tree=C10.586.250.520>C10.586.250 | ontology |
D000069578 | disease | is_a | D007239 | disease | mesh_tree | human | MeSH | tree=C01.936>C01 | ontology |
D000069584 | disease | is_a | D001943 | disease | mesh_tree | human | MeSH | tree=C04.588.180.800>C04.588.180|tree=C17.800.090.500.682>C17.800.090.500 | ontology |
D000069836 | disease | is_a | D017695 | disease | mesh_tree | human | MeSH | tree=C26.808.500>C26.808 | ontology |
D000069856 | disease | is_a | D007669 | disease | mesh_tree | human | MeSH | tree=C12.050.351.968.419.600.500.500>C12.050.351.968.419.600.500|tree=C12.050.351.968.967.249.500.500>C12.050.351.968.967.249.500|tree=C12.050.351.968.967.500.503.500>C12.050.351.968.967.500.503|tree=C12.200.777.419.600.500.500>C12.200.777.419.600.500|tree=C12.200.777.967.249.500.500>C12.200.777.967.249.500|tree=C12.20... | ontology |
D000070558 | disease | is_a | D005531 | disease | mesh_tree | human | MeSH | tree=C05.330.488.655>C05.330.488 | ontology |
D000070558 | disease | is_a | D005532 | disease | mesh_tree | human | MeSH | tree=C05.330.495.681>C05.330.495|tree=C05.660.585.512.380.813>C05.660.585.512.380|tree=C16.131.621.585.512.500.681>C16.131.621.585.512.500 | ontology |
D000070589 | disease | is_a | D000070558 | disease | mesh_tree | human | MeSH | tree=C05.330.488.655.500>C05.330.488.655|tree=C05.330.495.681.500>C05.330.495.681|tree=C05.660.585.512.380.813.500>C05.660.585.512.380.813|tree=C16.131.621.585.512.500.681.500>C16.131.621.585.512.500.681 | ontology |
D000070591 | disease | is_a | D005530 | disease | mesh_tree | human | MeSH | tree=C05.330.663>C05.330 | ontology |
D000070592 | disease | is_a | D005530 | disease | mesh_tree | human | MeSH | tree=C05.330.711>C05.330 | ontology |
D000070598 | disease | is_a | D007718 | disease | mesh_tree | human | MeSH | tree=C26.558.554.213>C26.558.554 | ontology |
D000070599 | disease | is_a | D014947 | disease | mesh_tree | human | MeSH | tree=C26.803>C26 | ontology |
D000070600 | disease | is_a | D007869 | disease | mesh_tree | human | MeSH | tree=C26.558.781>C26.558 | ontology |
D000070603 | disease | is_a | D001847 | disease | mesh_tree | human | MeSH | tree=C05.116.296>C05.116 | ontology |
D000070604 | disease | is_a | D005532 | disease | mesh_tree | human | MeSH | tree=C05.330.495.787>C05.330.495|tree=C05.660.585.512.380.875>C05.660.585.512.380|tree=C16.131.621.585.512.500.787>C16.131.621.585.512.500 | ontology |
D000070604 | disease | is_a | D013580 | disease | mesh_tree | human | MeSH | tree=C05.116.099.370.894.909>C05.116.099.370.894|tree=C05.660.906.909>C05.660.906|tree=C16.131.621.906.909>C16.131.621.906 | ontology |
D000070607 | disease | is_a | D009437 | disease | mesh_tree | human | MeSH | tree=C10.668.829.600.375>C10.668.829.600|tree=C23.888.592.612.664.275>C23.888.592.612.664 | ontology |
D000070607 | disease | is_a | D037061 | disease | mesh_tree | human | MeSH | tree=C05.360.500.500>C05.360.500|tree=C05.550.610.500>C05.550.610|tree=C23.888.592.612.540.500>C23.888.592.612.540 | ontology |
D000070617 | disease | is_a | D014947 | disease | mesh_tree | human | MeSH | tree=C26.599>C26 | ontology |
D000070624 | disease | is_a | D000070642 | disease | mesh_tree | human | MeSH | tree=C10.228.140.199.444.375>C10.228.140.199.444|tree=C10.900.300.087.235.375>C10.900.300.087.235|tree=C26.915.300.200.194.375>C26.915.300.200.194 | ontology |
D000070624 | disease | is_a | D003288 | disease | mesh_tree | human | MeSH | tree=C26.974.250.500>C26.974.250 | ontology |
D000070625 | disease | is_a | D001930 | disease | mesh_tree | human | MeSH | tree=C10.228.140.199.388>C10.228.140.199|tree=C10.900.300.087.219>C10.900.300.087|tree=C26.915.300.200.188>C26.915.300.200 | ontology |
D000070627 | disease | is_a | D000070642 | disease | mesh_tree | human | MeSH | tree=C10.228.140.199.444.500>C10.228.140.199.444|tree=C10.900.300.087.235.500>C10.900.300.087.235|tree=C26.915.300.200.194.500>C26.915.300.200.194 | ontology |
D000070627 | disease | is_a | D019636 | disease | mesh_tree | human | MeSH | tree=C10.574.250>C10.574 | ontology |
D000070627 | disease | is_a | D020208 | disease | mesh_tree | human | MeSH | tree=C10.228.140.199.500.500>C10.228.140.199.500|tree=C10.900.300.087.250.500>C10.900.300.087.250|tree=C23.550.291.500.063.500.500>C23.550.291.500.063.500|tree=C26.915.300.200.200.500>C26.915.300.200.200 | ontology |
D000070630 | disease | is_a | D004204 | disease | mesh_tree | human | MeSH | tree=C05.550.518.288>C05.550.518|tree=C26.289.288>C26.289 | ontology |
D000070631 | disease | is_a | D004204 | disease | mesh_tree | human | MeSH | tree=C05.550.518.192>C05.550.518|tree=C26.289.192>C26.289 | ontology |
D000070636 | disease | is_a | D000070599 | disease | mesh_tree | human | MeSH | tree=C26.803.063>C26.803 | ontology |
D000070636 | disease | is_a | D012421 | disease | mesh_tree | human | MeSH | tree=C26.761.340>C26.761 | ontology |
D000070636 | disease | is_a | D013708 | disease | mesh_tree | human | MeSH | tree=C26.874.400>C26.874 | ontology |
D000070639 | disease | is_a | D000092464 | disease | mesh_tree | human | MeSH | tree=C26.088.134.625>C26.088.134 | ontology |
D000070639 | disease | is_a | D052256 | disease | mesh_tree | human | MeSH | tree=C05.651.869.435>C05.651.869|tree=C26.874.800.500>C26.874.800 | ontology |
D000070642 | disease | is_a | D001930 | disease | mesh_tree | human | MeSH | tree=C10.228.140.199.444>C10.228.140.199|tree=C10.900.300.087.235>C10.900.300.087|tree=C26.915.300.200.194>C26.915.300.200 | ontology |
D000070656 | disease | is_a | D000070657 | disease | mesh_tree | human | MeSH | tree=C05.550.354.250>C05.550.354 | ontology |
D000070656 | disease | is_a | D002805 | disease | mesh_tree | human | MeSH | tree=C05.550.114.264.500>C05.550.114.264 | ontology |
D000070657 | disease | is_a | D007592 | disease | mesh_tree | human | MeSH | tree=C05.550.354>C05.550 | ontology |
D000070676 | disease | is_a | D052256 | disease | mesh_tree | human | MeSH | tree=C05.651.869.653>C05.651.869|tree=C26.874.800.750>C26.874.800 | ontology |
D000070779 | disease | is_a | D005870 | disease | mesh_tree | human | MeSH | tree=C04.557.450.565.380.690>C04.557.450.565.380 | ontology |
D000070779 | disease | is_a | D013585 | disease | mesh_tree | human | MeSH | tree=C05.550.870.445>C05.550.870 | ontology |
D000070779 | disease | is_a | D052256 | disease | mesh_tree | human | MeSH | tree=C05.651.869.762>C05.651.869 | ontology |
D000070896 | disease | is_a | D012784 | disease | mesh_tree | human | MeSH | tree=C26.404.625.500>C26.404.625|tree=C26.803.250.500>C26.803.250 | ontology |
MicrobeKG
MicrobeKG connects microorganisms, metabolites, substrates, diseases, host genes, and interventions in a heterogeneous knowledge graph. Records retain source and evidence fields for graph querying, resource analysis, and hypothesis generation.
This package contains the microbekg-audited graph: 3,647,004 assertion rows, 67,485 typed nodes, 25 relation labels, and 31 typed relation patterns. It is a lossless Parquet export prepared on 2026-09-29.
Data terms: the graph incorporates third-party sources with different terms.
The other label refers to source-specific terms, not a blanket open
license. See SOURCE_TERMS.md for source attribution, current
contribution counts, review dates, and unresolved redistribution permissions.
The software repository's MIT license does not license these third-party data.
Complete audited graph download
Download the complete audited graph ZIP from Google Drive.
This archive contains the complete audited graph (67,485 typed nodes and
3,647,004 assertions). It is separate from MicrobeKG-Bench_release_v3.zip,
which contains filtered graph triples, benchmark splits and associated tables.
Only the complete audited graph archive is currently available from this Drive link;
the benchmark v3 archive has not been uploaded there. The Drive archive may extract
to a differently named directory; pass that actual directory to --graph-dir.
Version names and graph scope
v3 / mkbench-v3 identifies the benchmark protocol and release package, not a
third revision of the complete knowledge graph. bench_v3/ contains its loader,
evaluator and release tools; analysis_v3/ contains the supplied research scripts.
The source graph is named microbekg-audited (local directory kg_build_audited):
67,485 typed nodes and 3,647,004 assertions. The separate v3 data archive contains
the same typed node identities and 2,671,916 graph triples, obtained by excluding
975,088 assertions supported only by CTD/HMDB. Its benchmark graph is therefore
a filtered derivative, not a byte-for-byte copy of the complete audited graph.
The archive also supplies benchmark splits, record tables, tested-null candidates
and an external test set. Hugging Face's existing nodes / edges configurations
remain the complete audited graph; they are not the v3 benchmark splits.
Date-free names are publication aliases. Source-review dates and temporal-cutoff
dates remain provenance facts, and random seeds remain unchanged. Archive hashes
and source hashes identify the exact payload. For an existing local graph folder
with another name, pass its actual path with --graph-dir; it need not be renamed.
MicrobeKG-Bench v3 — code update
The v3 benchmark provides study-aware, temporal and cold-endpoint evaluation,
17 internal settings across five seeds, and an external BugSigDB test set.
Its separate data package contains 2,671,916 triples after excluding CTD/HMDB-only
support. These are different from the 3,647,004 audited graph rows already hosted
under the edges configuration below.
- Code and analysis, with an installable loader, evaluator, CPU baselines and reference tests.
- Mirrored benchmark tools and release index.
- v3 data download: Google Drive release pending; a download URL has not yet
been provided. No new graph tables, benchmark splits or record tables are added
by this update. The existing
nodes/edgesconfigurations still load the microbekg-audited export.
After obtaining and extracting the data archive, use its v3/ directory:
git clone https://github.com/ZachGu-00/MicrobeKG-dataset_split_task.git
python -m pip install ./MicrobeKG-dataset_split_task/bench_v3
from microbekg_bench import load, Evaluator
from microbekg_bench.baselines import DegreePrior
case = load("t1_study_strict", 42, root="/path/to/v3")
result = Evaluator(case).evaluate(DegreePrior(case.train, case.n_entities))
print(result["metrics"])
Contents and loading
| Configuration | Split | Rows | Files |
|---|---|---|---|
edges |
full |
3,647,004 | 8 Parquet shards |
nodes |
full |
67,485 | 1 Parquet file |
full means the complete table. It is not a training or evaluation partition.
Shards preserve the original row order and contain at most 500,000 rows. The files
use Zstandard compression and row groups of at most 65,536 rows.
from datasets import load_dataset
repo_id = "ZachGu/MicrobeKG"
edges = load_dataset(repo_id, "edges", split="full")
nodes = load_dataset(repo_id, "nodes", split="full")
# Read progressively without materializing the entire table.
edge_stream = load_dataset(repo_id, "edges", split="full", streaming=True)
print(next(iter(edge_stream)))
For a private repository, first run hf auth login with an account that has access.
For reproducible work, pass revision="<dataset-commit-sha>" to load_dataset.
To use downloaded Parquet directly:
import pyarrow.dataset as ds
edges = ds.dataset("data/edges", format="parquet")
subset = edges.to_table(
columns=["head_id", "relation", "tail_id", "source", "evidence"],
filter=(ds.field("head_type") == "microbe")
& (ds.field("tail_type") == "disease"),
)
Schema
All columns are UTF-8 strings. Empty cells remain empty strings, and identifiers retain their original prefixes and formatting. See schema.json.
| Table | Column | Meaning |
|---|---|---|
| edges | head_id, head_type |
Identifier and type of the subject node |
| edges | relation |
Directed relation label |
| edges | tail_id, tail_type |
Identifier and type of the object node |
| edges | confidence |
Source-specific score or label, retained verbatim; not a calibrated probability |
| edges | species_source |
Source organism/context label, retained verbatim |
| edges | source |
Source labels; multiple labels can be separated by | |
| edges | evidence |
Source evidence, identifiers, and provenance, retained verbatim |
| edges | evidence_type |
Evidence-class labels, potentially combined with | |
| nodes | node_id |
Original canonical identifier |
| nodes | node_type |
One of the six entity types below |
| nodes | node_name |
Recorded display label; may be an identifier-derived label |
| nodes | source_databases |
Source labels associated with the node |
Node identity is (node_type, node_id). The same chemical identifier can occur
as both a substrate and a metabolite. Join edges to nodes using both the identifier
and type, rather than node_id alone. Evidence text may contain delimiters with
different meanings; it should not be interpreted as a single list of source labels.
| Node type | Count |
|---|---|
| metabolite | 25,366 |
| host_gene | 19,907 |
| microbe | 15,821 |
| disease | 5,212 |
| substrate | 1,057 |
| intervention | 122 |
The microbe count includes taxonomic ranks and genome bins; it is not a species count. The build audit flags identifier-derived display labels for 19,907 host genes, 252 metabolites, and 14 substrates. These cells are populated, not missing; the original labels and typed graph connections are retained without name imputation.
Sources and preparation
The graph integrates 18 upstream source labels, including curated association
databases, metabolic resources, taxonomy/ontology resources, and literature-derived
records. cross_source_conflict is an additional derived label. Source-labelled
counts overlap when a row cites multiple sources and should not be summed as
distinct graph assertions. Lit44 is a historical source identifier; this snapshot
contains retained assertions from 18 studies under that label.
The audited snapshot harmonizes typed identifiers and relation labels and preserves evidence and disagreement records. This export does not change, filter, rescore, or impute any graph field. Source TSV hashes, Parquet hashes, file sizes, and counts are recorded in manifest.json. Independently checked row equality, typed endpoint integrity, and statistics are recorded in validation.json and statistics.json.
Scope and appropriate use
Use the graph for evidence-aware retrieval, graph exploration, and development of research methods. Distinguish observed associations, computationally inferred metabolic capabilities, curated biochemical records, and ontology relations. Associations and graph paths alone do not establish causation or clinical efficacy. Source coverage, research attention, organism resolution, and evidence density are uneven; missing edges should not be assumed to be confirmed negative findings.
This package contains graph tables, not raw participant-level clinical records, sequencing reads, upstream database dumps, model checkpoints, or benchmark splits. Previously reported benchmark results use the frozen reference graph and splits; they are not new measurements on this audited export. Define and document suitable splits and leakage controls when evaluating methods on this snapshot.
Attribution and related code
Code: MicrobeKG-dataset_split_task. The code repository documents the scope of its reference graph construction and evaluation tools; this export is not a claim of complete upstream reconstruction.
When citing the resource, include MicrobeKG, microbekg-audited, the actual Hugging Face repository URL, and the immutable dataset commit used. Also acknowledge the relevant original data providers listed in SOURCE_TERMS.md. No DOI or publication identifier has been assigned by this packaging operation. Questions about this package can be filed in the linked code repository's Issues.
- Downloads last month
- 223