BioXArena-Pro Public
Public data package of BioXArena-Pro (release v5.1.1, 2026-10-21): 64 harder biology tasks in 8 domains, 8 tasks per domain.
Each task folder contains public/description.md (task, data construction, split, metric and rules),
public/sample_submission.csv, the training and test inputs, and any modality-specific assets. Test labels are not
included; they are in the access-controlled repository mbzuai-ai4bio/BioXArena-Pro-Private.
Every task was audited for label leakage, shortcuts and description errors before release; the per-task
description.md records the data version and what changed. Scores obtained on earlier internal versions are not
comparable with this release.
Changes in v5.1.1 (2026-10-21)
A final check of all 64 tasks found two tasks whose test information could be turned into score; both were fixed:
phenotype-disease/covid-nb-hospitalization-forecast(data v5.1): the test set now has two forecast origins (weeks 76 and 103; 100 records) instead of four. With origins 9 weeks apart, a later origin's record allowed the earlier origin's target-week rates to be reconstructed. Same records and labels for the kept origins.single-cell-spatial/gastrulation-xl-kinetic-state(v5.2.1, description only): a sentence that stated the stages of the test-only embryo pools was removed. Data unchanged.
Description corrections without data changes (removed test-set counts or corrected numbers and wording): boom-xl-csd, tdc-xl-dti-binding-affinity, pring-xl-cross-species-ppi, lnctard-xl-relation, cellverse-xl-cta, tahoe-xl-single-cell, clinvar-nb-disease-variant-effect, skempi-xl-ppi-ddg, asep-xl-epitope, subcell-xl-localization, calcium-nb-neuron-segmentation, medmnist3d-xl-decathlon.
Layout
BioXArena-Pro-Data-Public/
└── <domain>/
└── <task>/
└── public/
├── description.md
├── sample_submission.csv
├── training data
├── test data
└── modality-specific assets
Download
Full package (extracts to BioXArena-Pro-Data-Public/<domain>/<task>/public/):
hf download mbzuai-ai4bio/BioXArena-Pro-Public BioXArena-Pro-Data-Public.tar.gz --repo-type dataset --local-dir .
tar -xzf BioXArena-Pro-Data-Public.tar.gz
Size: 101.73 GB (sha256 72bef86dc7c0e6dffd09affb3aec33b7af20590dbda400042b2c58788c76f0a9).
One archive per domain under Data-Public-by-Domain/ (each extracts to <domain>/<task>/public/);
Data-Public-by-Domain/MANIFEST.txt lists the tasks and checksums.
| Archive | Size | # Tasks | Tasks Included |
|---|---|---|---|
Data-Public-by-Domain/chemical-biology-8tasks-all.tar.gz |
305.54 MB | 8 | All 8 chemical-biology tasks |
Data-Public-by-Domain/imaging-8tasks-all.tar.gz |
57.63 GB | 8 | All 8 imaging tasks |
Data-Public-by-Domain/network-biology-8tasks-all.tar.gz |
3.69 GB | 8 | All 8 network-biology tasks |
Data-Public-by-Domain/perturbation-8tasks-all.tar.gz |
852.27 MB | 8 | All 8 perturbation tasks |
Data-Public-by-Domain/phenotype-disease-8tasks-all.tar.gz |
313.71 MB | 8 | All 8 phenotype-disease tasks |
Data-Public-by-Domain/sequence-genomics-8tasks-all.tar.gz |
7.25 GB | 8 | All 8 sequence-genomics tasks |
Data-Public-by-Domain/single-cell-spatial-8tasks-all.tar.gz |
3.81 GB | 8 | All 8 single-cell-spatial tasks |
Data-Public-by-Domain/structure-8tasks-all.tar.gz |
27.87 GB | 8 | All 8 structure tasks |
Tasks
| Domain | Task | Title |
|---|---|---|
chemical-biology |
boom-xl-csd |
BOOM-XL-CSD: Out-of-Distribution Molecular Property Prediction |
chemical-biology |
chemixhub-xl-conductivity |
CheMixHub-XL Conductivity: Mixture Ionic Conductivity |
chemical-biology |
drugcomb-xl-multiomics-synergy |
DrugComb-XL: Synergy of New Compounds in NCI-60 Combination Screens |
chemical-biology |
kano-molecular-property-prediction |
Molecular Property Prediction (KANO benchmark suite) |
chemical-biology |
massspecgym-xl-msms-retrieval |
MassSpecGym-XL: Molecule Retrieval from Tandem Mass Spectra |
chemical-biology |
solubility-xl |
Solubility-XL: Aqueous Solubility Prediction |
chemical-biology |
tdc-xl-dti-binding-affinity |
TDC-XL-DTI-Binding-Affinity: Drug–Target Affinity Prediction |
chemical-biology |
uspto-nb-reaction-product |
Organic Reaction Product Prediction on New Substrates (USPTO-480k) |
imaging |
calcium-nb-neuron-segmentation |
Neuron Segmentation from Simulated and Real Two-Photon Calcium Imaging |
imaging |
cryocrab-xl-quality |
CryoCRAB-XL: Cryo-EM Micrograph Quality (CTF Fit Resolution) Regression |
imaging |
drugclass-xl-if |
DrugClass-XL: Drug-Treatment Classification in a New Screen |
imaging |
mammovqa-xl-external |
MammoVQA-XL: Mammography Visual Question Answering |
imaging |
medmnist3d-xl-decathlon |
MedMNIST3D-XL: Classification Under an Unknown Acquisition Shift (five sources) |
imaging |
microscopy-instance-segmentation |
Microscopy Instance Segmentation |
imaging |
rocov2-xl-retrieval |
ROCOv2-XL: Matching a Radiology Image to Its Caption Among Concept-Similar Candidates |
imaging |
subcell-xl-localization |
SubCell-XL: Subcellular Localization of Tagged Proteins from Microscopy Crops |
network-biology |
causalbench-xl-grn |
CausalBench-XL: Which Perturbations Move a Gene |
network-biology |
lnctard-xl-relation |
LncTarD-XL: Relation-Conditioned Target Ranking |
network-biology |
music-xl-seccomplex |
MuSIC-XL-SECComplex: Protein Co-Complex Discovery from SEC-MS |
network-biology |
ppi-nb-cross-species |
Protein-Protein Interaction Prediction |
network-biology |
pring-xl-cross-species-ppi |
PRING-XL: Cross-Species Protein-Protein Interaction Prediction |
network-biology |
scigym-xl-sbml |
SciGym-XL: Completing Partial Pathway Models of Unseen Pathways (SBML) |
network-biology |
swissprot-nb-protein-function |
Protein Function Prediction from Sequence |
network-biology |
synlethdb-xl-sl |
SynLethDB-XL: Synthetic-Lethal Partner Ranking |
perturbation |
cellverse-xl-cta |
CellVerse-XL: annotate an MS cohort with a healthy-control reference |
perturbation |
cpjump-xl-genetic-matching |
CPJUMP-XL: Cold-Target Compound-to-Gene Matching from Cell Painting |
perturbation |
gdsc-xl-lineage-holdout-response |
GDSC-XL: Drug Response in Cancer Lineages Not Seen in Training |
perturbation |
perturbench-xl-response |
PerturBench-XL: Context-Specific Perturbation Response |
perturbation |
scgenescope-xl-treatment |
scGeneScope-XL: Single-Cell Treatment Classification |
perturbation |
sciplex-xl-distribution |
sci-Plex-XL: Transcriptional Response of Active Drugs in Unseen Drug × Cell-Line Pairs |
perturbation |
tahoe-xl-single-cell |
Tahoe-100M: drug response of a new cell line from its baseline profile |
perturbation |
xatlas-orion-xl-perturb |
X-Atlas/Orion-XL: Transcriptional Response to the Knock-down of an Unseen Gene |
phenotype-disease |
clinvar-nb-disease-variant-effect |
Disease-Specific Variant Effect Prediction |
phenotype-disease |
covid-nb-hospitalization-forecast |
COVID-19 Hospitalization Trend Forecasting (multi-origin) |
phenotype-disease |
cptac-pan-xl-protein-imputation |
CPTAC-Pan-XL: External Protein Imputation |
phenotype-disease |
gutcrc-xl-cross-cohort-detection |
GutCRC-XL: Detecting Colorectal Cancer from Stool Metagenomes in Unseen Cohorts |
phenotype-disease |
msk-met-xl-tumor-origin |
MSK-MET-XL: Assay-Generation-OOD Tumor Origin |
phenotype-disease |
survpath-xl-multimodal |
SurvPath-XL-Multimodal: TCGA survival from histology thumbnails + bulk RNA-seq |
phenotype-disease |
target-aml-xl-cytogenetics |
TARGET-AML-XL: Protocol-OOD cytogenetic risk-group classification from clinical + multi-omics data |
phenotype-disease |
tcga-pancan-xl-multiomics-survival |
TCGA-PanCan-XL: pan-cancer overall survival from three de-identified omics layers |
sequence-genomics |
aso-xl-knockdown-efficacy |
ASO-XL: Antisense Oligonucleotide Knockdown Efficacy |
sequence-genomics |
cgbench-xl-variant |
CGBench-XL: Clinical Variant Assertion from Variant Context |
sequence-genomics |
de-novo-peptide-sequencing-from-tandem-mass-spectrometry |
De Novo Peptide Sequencing from Tandem Mass Spectrometry |
sequence-genomics |
deepstarr-nb-genomic-sequence |
Genomic Sequence Prediction (14 instances) |
sequence-genomics |
escape-xl-amp-multilabel |
ESCAPE-XL: multilabel antimicrobial-peptide classification |
sequence-genomics |
proteingym-xl-fitness |
ProteinGym-XL: Region-Held-Out Mutation Fitness |
sequence-genomics |
sirna-xl-knockdown-efficacy |
siRNA-XL: RNA Interference Knockdown Efficacy |
sequence-genomics |
yeast-nb-promoter-expression |
Yeast Promoter Expression Prediction |
single-cell-spatial |
dlpfc-xl-spatial-domain |
DLPFC-XL: Donor-OOD Spatial-Domain Annotation |
single-cell-spatial |
gastrulation-xl-kinetic-state |
Gastrulation-XL: Developmental Stage and Cell Fate of Held-out Mouse Embryos |
single-cell-spatial |
hest-coad-xl-spatial-expression |
HEST-COAD-XL: Patient-OOD Spatial Expression |
single-cell-spatial |
lung-cite-xl-cell-matching |
CITE-XL: Cross-Protocol Cell Matching in Human PBMC CITE-seq |
single-cell-spatial |
openproblems-xl-rna-atac |
OpenProblems-XL RNA-to-ATAC: Donor-OOD Chromatin Accessibility of Cell States |
single-cell-spatial |
openproblems-xl-rna-protein |
OpenProblems-XL RNA-to-Protein: Donor-OOD CITE-seq Translation |
single-cell-spatial |
seqfish-xl-deconvolution |
Hippocampus-XL: Spatial Deconvolution into 16 Cell Types |
single-cell-spatial |
tabula-sapiens-xl-ood-annotation |
Tabula-Sapiens-XL: Donor-OOD Immune and Haematopoietic Cell Annotation |
structure |
asep-xl-epitope |
Antibody-Specific Epitope Residues |
structure |
cath-nb-inverse-folding |
Inverse Protein Folding (CATH 4.2 / 4.3, PDB2022) |
structure |
cpsea-xl-cyclic-peptide-affinity |
Cyclic Peptide–Protein Binding Energy from Coarse Structure |
structure |
decoydb-xl-pose-classify |
DecoyDB-XL-PoseClassify: Ranking Docked Poses Inside the Known Binding Site |
structure |
denoise-cryo-xl |
DenoiseCryo-XL: Backbone Tracing and Residue Assignment in Experimental Cryo-EM Maps |
structure |
ppi-nb-interface-residues |
Partner-specific protein-protein interface residue prediction |
structure |
psbench-xl-multimer-mqa |
Quality Assessment of Predicted Protein-Complex Structures — Small Model Pools |
structure |
skempi-xl-ppi-ddg |
Protein-protein binding free-energy change upon mutation (SKEMPI 2.0 ddG) |
Scoring
Every grader returns a higher-is-better score in [0, 1]. Correlation and MCC metrics are reported as clip(r, 0, 1)
(chance = 0); AUC, c-index, AP, F1, accuracy and MRR keep their native scale (for example chance = 0.5 for AUC).
tahoe and sciplex are offset by the all-zero prediction, covid uses an RPS skill score. A task's
public/description.md documents its inputs, split, metric and rules; the graders are in the BioXArena-Pro code
release (tasks/<domain>/<task>/grade.py).
Licence
The packaging, splits, opaque identifiers and task descriptions are released under CC BY-NC 4.0. The underlying
data come from public resources that keep their own licences and citation requirements; each task's
public/description.md names its source.
Contact
MBZUAI AI4Bio (mbzuai-ai4bio on Hugging Face). Please report data problems in the repository's Community tab.
- Downloads last month
- 47