TACK — TArgeting Chimeras Knowledge

A curated, ML-ready dataset and benchmark for PROTAC degradation activity prediction

---

Overview

Proteolysis-targeting chimeras (PROTACs) are a promising drug modality that induces targeted protein degradation by hijacking the cell's native ubiquitin–proteasome system. However, rational PROTAC design remains challenging due to the complex interplay between molecular structure, target proteins (POIs), E3 ligases, and cellular context.

TACK addresses three critical gaps in existing PROTAC ML benchmarks:

• Data scarcity and inconsistency — harmonized from three major public repositories with standardized SMILES, protein annotations, and experimental conditions
• Lack of rigorous benchmarks — scaffold-based 5×5 cross-validation and formal statistical testing (Friedman, Wilcoxon, Tukey HSD) rather than simple train/test splits
• Limited scope — supports regression (pDC₅₀, Dmax) and binary classification, not just the binary classification framing common in prior work

Ribes*, Dunlop*, and Mercado. TACK: A statistical evaluation of degradation activity on a novel TArgeting Chimeras Knowledge dataset. KDD AI for Sciences Track, 2026.

---

Dataset at a Glance

Statistic	TPDdb	PROTAC-DB	PROTACpedia	TACK
Total records	22,183	9,380	1,203	6,561
Unique PROTACs	21,429	6,110	1,189	3,514
Degradation endpoints	6,518	2,170	580	6,561
POI targets	184	441	79	164
E3 ligases	8	21	8	9
Cell lines	142	—	139	155
Hold-out set entries	—	—	—	913

The dataset contains 4,184 DC₅₀ and 2,377 Dmax measurements. Approximately 55% of entries are classified as active (DC₅₀ ≤ 100 nM and Dmax ≥ 80%).

Most represented biology:

• POIs: Androgen Receptor (26.7%), SMARCA2 (12.0%), BTK (8.9%) — top 5 account for 61.6% of endpoints
• E3 ligases: CRBN and VHL together account for 98.7% of measurements
• Cell lines: LNCaP (20.9%), SW1573 (15.2%), Mino (7.0%) across 155 unique lines

---

Configs (Subsets)

Config	Description	Examples
default	All endpoints (DC₅₀ + Dmax combined)	6,561
DC50	Potency measurements only	4,184
Dmax	Maximal degradation efficacy only	2,377
multitask	Paired DC₅₀ + Dmax for the same PROTAC/assay, also used for binary activity classification	1,563

---

Loading the Dataset

from datasets import load_dataset

# All endpoints
ds = load_dataset("ailab-bio/TACK")

# DC50 only
ds_dc50 = load_dataset("ailab-bio/TACK", "DC50")

# Dmax only
ds_dmax = load_dataset("ailab-bio/TACK", "Dmax")

# Paired multitask (DC50 + Dmax for same compound)
ds_multi = load_dataset("ailab-bio/TACK", "multitask")

The SMILES_Held_Out column flags the structurally dissimilar held-out set (~10% of data). To reproduce the train/validation split used in the paper:

ds = load_dataset("ailab-bio/TACK", "DC50")["train"]
train_val = ds.filter(lambda x: not x["SMILES_Held_Out"])
held_out  = ds.filter(lambda x: x["SMILES_Held_Out"])

---

Schema

Common columns (all configs)

Column	Type	Description
SMILES	string	Canonical RDKit SMILES of the PROTAC
POI_Name	string	Gene name of the protein of interest
POI_Sequence	string	UniProt amino acid sequence of the POI
POI_UniProt	string	UniProt accession for the POI
Ligase_Name	string	Name of the E3 ligase
Ligase_Sequence	string	UniProt amino acid sequence of the E3 ligase
Ligase_UniProt	string	UniProt accession for the E3 ligase
Cell_Line	string	Cell line name (Cellosaurus-standardized)
Cell_Line_ID	string	Cellosaurus identifier
Cell_Line_Species	string	Species of origin for the cell line
Cell_Line_Description	string	Textual description from Cellosaurus
Value	float64	Measured endpoint value
Value_Unit	string	Unit of the value (nM for DC₅₀; % for Dmax)
Value_Type	string	Endpoint type (DC50 or Dmax)
Value_Operator	string	Comparison operator if censored (<, >, etc.)
Value_Error	float64	Reported measurement error (if available)
Value_Range_Min	float64	Lower bound for range-reported values
Value_Range_Max	float64	Upper bound for range-reported values
Value_Concentration	float64	Treatment concentration for Dmax assays
Value_Concentration_Unit	string	Unit of treatment concentration
Value_Category	string	Categorical activity label (where applicable)
Assay_Time	float64	Treatment duration in hours
Assay	string	Standardized assay type (e.g., Western Blot, HiBit)
Modality	string	Compound modality (PROTAC, molecular glue, etc.)
Reference	string	Source literature or patent reference
Description	string	Original assay description
Database	string	Source database (TPDdb, PROTAC-DB, PROTACpedia)
TPD_ID	string	Original identifier in the source database
POI_Cluster	int64	POI sequence-based cluster ID
SMILES_Scaffold_Cluster	int64	Murcko scaffold cluster ID (used for CV splitting)
SMILES_Butina_Cluster	int64	Butina/Tanimoto cluster ID
SMILES_Held_Out	bool	True if this entry belongs to the structural hold-out set

The multitask config replaces Value, Value_Unit, etc. with Value_DC50, Value_Dmax and analogous suffixed columns, enabling simultaneous supervision on both endpoints.

---

Curation Pipeline

Raw data was aggregated from three repositories and cleaned through a multi-stage pipeline:

1. SMILES standardization — canonicalization via RDKit; duplicates resolved by weighted scoring
2. Endpoint standardization — DC₅₀ values converted to nM; range values converted to arithmetic mean with bounds stored; censored values (operators) preserved but flagged for exclusion from evaluation sets; categorical patent grades excluded
3. Protein annotation — POI and E3 ligase names mapped to UniProt accessions; amino acid sequences retrieved from UniProt; BRD4 isoform handling applied
4. Cell line standardization — validated against Cellosaurus; standardized identifiers and descriptions attached
5. Assay standardization — assay descriptions parsed and normalized (e.g., "WB" → "Western Blot"); treatment concentrations extracted from metadata and free-text fields
6. Hold-out construction — ~10% most structurally dissimilar PROTACs (by average Tanimoto distance using 512-bit Morgan8 fingerprints) isolated before any model training
7. Scaffold-based CV clustering — remaining data grouped by Murcko scaffolds to prevent leakage between CV folds

Please refer to this link for the codebase used to generate the data.

---

Benchmark Results

Models were evaluated using scaffold-based 5×5 repeated cross-validation with statistical testing (Friedman + Wilcoxon + Benjamini–Hochberg correction for feature selection; Tukey HSD for architecture comparison).

Regression (hold-out set)

Task	Model	MAE	RMSE	R²	Spearman ρ
pDC₅₀	MLP	0.58	0.80	0.66	0.76
Dmax	XGBoost	18.86	25.68	0.36	0.66

Binary Classification (validation folds, mean ± CI)

Model	ROC-AUC	PR-AUC	MCC	Recall @ Prec≥0.8
XGBoost	0.851	0.870	0.523	0.777
MLP	0.799	0.824	0.448	0.600
PROTAC-STAN	0.746	0.773	0.405	0.443

All pairwise differences are statistically significant (Tukey HSD, p < 0.001).

Key findings:

• pDC₅₀ is substantially more predictable than Dmax (R² 0.66 vs. 0.36), reflecting the greater dependence of maximal degradation on cellular factors not captured by current molecular representations
• Classical tree-based methods (XGBoost) outperform the domain-specific GNN PROTAC-STAN, consistent with the small-to-medium tabular dataset regime
• Simple one-hot or n-gram protein encodings often match expensive ESM-S embeddings, particularly with XGBoost; ESM-S embeddings give a modest advantage for XGBoost on the classification task

Ensemble Uncertainty Quantification

Ensemble standard deviation correlates positively with absolute prediction error, enabling confidence-aware compound prioritization:

Task	Method	RMSE	Spearman ρ (σ vs. |error|)
pDC₅₀	Caruana (33 models)	0.672	0.207*
pDC₅₀	Uniform avg (500 models)	0.683	0.355**
Dmax	Caruana (22 models)	21.32	0.543**
Dmax	Uniform avg (500 models)	22.92	0.694**

** p < 0.001; * p < 0.01

---

Best Models

Pre-trained ensemble models (XGBoost and MLP, all feature configurations, 25 CV folds each) are available on Zenodo:

---

Citation

If you use TACK in your research, please cite:

@misc{ribes2026tackstatisticalevaluationdegradation,
      title={TACK: A statistical evaluation of degradation activity on a novel TArgeting Chimeras Knowledge dataset}, 
      author={Stefano Ribes and Nils Dunlop and Rocío Mercado},
      year={2026},
      eprint={2605.19579},
      archivePrefix={arXiv},
      primaryClass={q-bio.QM},
      url={https://arxiv.org/abs/2605.19579}, 
}

---

License

The TACK dataset is released under the MIT License. The underlying experimental data is sourced from PROTAC-DB, PROTACpedia, and TPDdb — please refer to the respective database licenses for conditions on downstream use of the original data.

---

Acknowledgements

SR and RM acknowledge funding from the Chalmers Gender Initiative for Excellence (Genie). RM and ND acknowledge funding from the Wallenberg AI, Autonomous Systems and Software Program (WASP), supported by the Knut and Alice Wallenberg Foundation. The authors thank Yossra Gharbi, Alexander Persson, and Felix Erngård for helpful discussions. Computations were enabled by Chalmers e-Commons and NAISS (grant 2022-06725).