# The Well for The Cell: 139,039 whole-cell E. coli simulation trajectories from the wcEcoli model

## Facts

- Provider: Alexey Yermakov (cSynBioSys); wcEcoli model by the Covert Lab, Stanford
- License: CC-BY-4.0 (https://huggingface.co/datasets/cSynBioSys/well4cell)
- Size: 978 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/cSynBioSys/well4cell
- Access: public, readable without an account
- Category: Life Sciences
- Source: https://huggingface.co/datasets/cSynBioSys/well4cell
- Page: https://scigantic.com/archive/0b4aaf1e-4bdf-56ef-9e4b-2533cfb644cd
- Keywords: Life Sciences, time series, Parquet, Hugging Face

## About

Time series from the Covert lab wcEcoli whole-cell model of Escherichia coli, assembled by Alexey Yermakov. Each sample is one cell generation, from birth to division, recorded every 1 s of simulated time, and generations of a lineage are linked by parent_id. The 139,039 trajectories come from 36,828 lineages of 4,618 distinct cells in five families: wild type in static media (1,281), single-gene knockouts across media (97,478), dynamic media shifts and sinusoidal schedules (21,404), transcription-factor and ppGpp set-point perturbations (2,046) and double knockouts (16,830). The card documents limitations as index columns rather than removing rows: partial knockouts of genes in several transcription units, collateral genes silenced, bit-identical duplicate designs, glucose doses that act as one condition, and cells that divide after translation collapses. Layout: index.parquet, manifest.json, data/, provenance/ and stats/, with lineage- and TU-held-out splits; about 1.67 TB of Parquet. Useful for virtual cell and surrogate models of metabolism and gene expression. Everything in the repo is CC BY 4.0; the wcEcoli simulator itself is under a separate Stanford academic license and is not included.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("cSynBioSys/well4cell")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/charging/double_knockout/part-00000.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/0b4aaf1e-4bdf-56ef-9e4b-2533cfb644cd
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/0b4aaf1e-4bdf-56ef-9e4b-2533cfb644cd
- Over MCP: `get_schema_card` with id `0b4aaf1e-4bdf-56ef-9e4b-2533cfb644cd` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Time series datasets](https://scigantic.com/topics/time-series)
- [Life Sciences datasets](https://scigantic.com/topics/life-sciences)
