SciganticGuidesTopicsExplore

How to open the OpenADMET CYP inhibition challenge data in a Jupyter notebook

The dataset is five CSV files, and pandas reads each one directly: pd.read_csv(mount / "cyp-challenge-TRAIN_inhibition.csv") gives the 4,905-compound training table. In a Scigantic notebook the Hugging Face repository openadmet/cyp-challenge-train-test is mounted at /mnt/archive, so no datasets download or cache is involved.

The code on this page runs in a Scigantic hosted notebook, where the dataset is mounted read-only at /mnt/archive. The same path is in the SCIGANTIC_MOUNT_PATH environment variable and linked from ~/archive. The public source is the Hugging Face dataset openadmet/cyp-challenge-train-test (public).

Open OpenADMET CYP challenge data on SciganticOpen the sample notebook: A scored blind-challenge entry, rebuilt

Code

Read the training set and summarize one isoform

# Start here: click this cell, press Shift+Enter.
import os
from pathlib import Path
import pandas as pd

ISOFORM = "CYP3A4"          # try CYP1A2, CYP2C9 or CYP2D6, then run again

mount = Path(os.environ.get("SCIGANTIC_MOUNT_PATH") or Path.home() / "archive")
train = pd.read_csv(mount / "cyp-challenge-TRAIN_inhibition.csv")

col = f"{ISOFORM}_pIC50_direct_inhibition"
d = train[[col, f"{col}_conf_low", f"{col}_conf_high"]].dropna()
width = d[f"{col}_conf_high"] - d[f"{col}_conf_low"]

print(f"{len(train):,} training compounds, {len(d):,} of them with a measured {ISOFORM} pIC50")
print(f"median pIC50 {d[col].median():.2f}, strongest inhibitor {d[col].max():.2f}")
print(f"each label carries a 95% credible interval, {width.median():.2f} log units wide at the median")
print(f"{(width > 1).mean() * 100:.0f}% of them are wider than a whole log unit")
print()
print("A prediction anywhere inside a compound's interval scores zero error.")
print("The challenge is graded on that, and the recipe below is built around it.")

Copied from this dataset's starter notebook on Scigantic, the one that opens when you start a notebook from the archive page (one cell, unchanged).

What the files are

Training and blinded test data for the OpenADMET CYP Inhibition Blind Challenge, covering direct and time-dependent inhibition of four cytochrome P450 isoforms: CYP1A2, CYP2C9, CYP2D6 and CYP3A4.

Each training label is a pIC50 with a 95% credible interval from the dose-response fit. The label matrix is sparse: 3,596 of the 4,905 compounds have a value for only one of the four isoforms.

How the files are laid out

Path under the mountWhat it holds
cyp-challenge-TRAIN_inhibition.csv4,905 rows by 18 columns. Direct-inhibition pIC50 for the four isoforms, with interval bounds. The regression target.
cyp-challenge-TEST-BLINDED.csv750 rows by 2 columns: Molecule_Name and SMILES.
cyp-challenge-TRAIN_TDI.csv6,145 rows by 36 columns. is_TDI labels for CYP2D6 and CYP3A4 plus pIC50 from both assay arms.
cyp-challenge-single-concentration-TRAIN.csv17,504 rows by 12 columns. The primary screen in long format, one row per compound and enzyme.
cyp-challenge-TRAIN_Emax.csv6,146 rows by 30 columns. Maximal effect plus is_TDI for all four isoforms.

Size, format and license

FormatCSV
Size5 files, 6.7 MB. 6,146 training compounds and 750 blinded test compounds.
LicenseApache 2.0
ProviderOpenADMET
Public sourcethe Hugging Face dataset openadmet/cyp-challenge-train-test (public)
Notebook imageCryo-EM and structural biology (RDKit and pandas preinstalled)

Things to know before you read it

Run it

Start a notebook from the OpenADMET CYP challenge data archive page. The data is mounted when the notebook opens, so the code above runs as written and nothing is downloaded first. For a longer worked example, open the sample notebook A scored blind-challenge entry, rebuilt.

Related guides

Last checked 2026-11-21. All dataset guides.