README.md

---
license: mit
language:
- en
tags:
- biology
- genomics
- yeast
- transcription-factors
- gene-expression
- perturbation-screen
- overexpression
- knockout
- microarray
- functional-genomics
pretty_name: "Hughes 2006 Yeast Transcription Factor Perturbation Dataset"
size_categories:
- 100K<n<1M
configs:
- config_name: metadata
  description: Transcription factor metadata including essentiality and QC status
  dataset_type: metadata
  default: true
  applies_to: ["overexpression", "knockout"]
  data_files:
  - split: train
    path: metadata.parquet
  dataset_info:
    features:
    - name: sample_id
      dtype: integer
      description: >-
        unique identifier for a specific sample. The sample ID identifies
        a unique regulator_locus_tag and can be used to join to the
        other datasets in this repo, including the metadata
    - name: regulator_locus_tag
      dtype: string
      role: identifier
      description: >-
        Systematic gene name (ORF identifier) of the
        transcription factor
    - name: regulator_symbol
      dtype: string
      description: Standard gene symbol of the transcription factor
    - name: found_domain
      dtype: string
      description: >-
        Identified DNA-binding domain(s) or protein family classification
    - name: sgd_description
      dtype: string
      description: >-
        Functional description from Saccharomyces Genome Database (SGD)
    - name: essential
      dtype: bool
      description: >-
        Boolean indicating whether the gene is essential for viability
    - name: oe_passed_qc
      dtype: bool
      description: >-
        Boolean indicating whether overexpression experiments passed
        quality control
    - name: del_passed_qc
      dtype: bool
      description: >-
        Boolean indicating whether deletion experiments passed
        quality control

- config_name: overexpression
  description: Overexpression perturbation normalized log2 fold changes
  dataset_type: annotated_features
  data_files:
  - split: train 
    path: overexpression.parquet
  experimental_conditions:
    temperature_celsius: unspecified
    cultivation_method: unspecified
    media:
      # Hughes et al 2006: "selective medium supplemented with 2% raffinose"
      name: selective_medium
      carbon_source:
        - compound: D-raffinose
          # Hughes et al 2006: 2% raffinose
          concentration_percent: 2
      nitrogen_source: unspecified
    induction:
      # Hughes et al 2006: "induction with 2% galactose for 3 h"
      inducer:
        compound: D-galactose
        concentration_percent: 2
      duration_hours: 3
  dataset_info:
    features:
    - name: sample_id
      dtype: integer
      description: >-
        unique identifier for a specific sample. The sample ID identifies
        a unique regulator_locus_tag and can be used to join to the
        other datasets in this repo, including the metadata
    - name: regulator_locus_tag
      dtype: string
      description: >-
        Systematic gene name (ORF identifier) of the
        perturbed transcription factor
      role: regulator_identifier
    - name: regulator_symbol
      dtype: string
      description: Standard gene symbol of the perturbed transcription factor
    - name: target_locus_tag
      dtype: string
      description: >-
        Systematic gene name (ORF identifier) of the
        target gene measured
      role: target_identifier
    - name: target_symbol
      dtype: string
      description: Standard gene symbol of the target gene measured
      role: target_identifier
    - name: dye_plus
      dtype: float64
      role: quantitative_measure
      description: >- 
        Normalized log2 fold change for positive (+) dye orientation.
        Positive values indicate upregulation in response to overexpression.
    - name: dye_minus
      dtype: float64
      role: quantitative_measure
      description: >-
        Normalized log2 fold change for negative (-) dye orientation.
        Positive values indicate upregulation in response to overexpression.
    - name: mean_norm_log2fc
      dtype: float64
      role: quantitative_measure
      description: >-
        Average log2 fold change across dye orientations,
        providing a dye-independent estimate of gene expression
        change upon transcription factor overexpression.
    - name: responsive
      dtype: float64
      description: >-
        abs(mean_norm_log2fc) `>` 1. Note that the authors do not have a threshold
        in their paper. They do use a zscore `>` 3 which they say is about a 
        1.58 FC. That yielded exceptionally few DE genes, so I reduced the value

- config_name: knockout
  description: Deletion/knockout perturbation normalized log2 fold changes
  dataset_type: annotated_features
  data_files:
  - split: train
    path: knockout.parquet
  experimental_conditions:
    temperature_celsius: unspecified
    cultivation_method: unspecified
    media:
      # Hughes et al 2006: "synthetic medium supplemented with 2% dextrose"
      name: synthetic_medium
      carbon_source:
        - compound: D-glucose
          # Hughes et al 2006: 2% dextrose
          concentration_percent: 2
      nitrogen_source: unspecified
  dataset_info:
    features:
    - name: sample_id
      dtype: integer
      description: >-
        unique identifier for a specific sample. The sample ID identifies
        a unique regulator_locus_tag and can be used to join to the
        other datasets in this repo, including the metadata
    - name: regulator_locus_tag
      dtype: string
      description: >-
        Systematic gene name (ORF identifier) of the perturbed
        transcription factor
      role: regulator_identifier
    - name: regulator_symbol
      dtype: string
      description: Standard gene symbol of the perturbed transcription factor
      role: regulator_identifier
    - name: target_locus_tag
      dtype: string
      description: >-
        Systematic gene name (ORF identifier) of the
        target gene measured
      role: target_identifier
    - name: target_symbol
      dtype: string
      description: Standard gene symbol of the target gene measured
      role: target_identifier
    - name: dye_plus
      dtype: float64
      description: >-
        Normalized log2 fold change for positive (+) dye orientation.
        Positive values indicate upregulation in response to deletion.
      role: quantitative_measure
    - name: dye_minus
      dtype: float64
      description: >-
        Normalized log2 fold change for negative (-) dye orientation.
        Positive values indicate upregulation in response to deletion.
      role: quantitative_measure
    - name: mean_norm_log2fc
      dtype: float64
      description: >-
        Average log2 fold change across dye orientations, providing a
        dye-independent estimate of gene expression change upon
        transcription factor deletion.
      role: quantitative_measure
    - name: responsive
      dtype: float64
      description: >-
        abs(mean_norm_log2fc) `>` 1. Note that the authors do not have a threshold
        in their paper. They do use a zscore `>` 3 which they say is about a 
        1.58 FC. That yielded exceptionally few DE genes, so I reduced the value
---
# Hughes 2006

This data is parsed from data presented in

[G. Chua, Q.D. Morris, R. Sopko, M.D. Robinson, O. Ryan, E.T. Chan, B.J. Frey, B.J.
Andrews, C. Boone, & T.R. Hughes, Identifying transcription factor functions and targets
by phenotypic activation, Proc. Natl. Acad. Sci. U.S.A. 103 (32) 12045-12050,
https://doi.org/10.1073/pnas.0605140103
(2006).](https://doi.org/10.1073/pnas.0605140103)

The data is made [available by the
author](https://hugheslab.ccbr.utoronto.ca/supplementary-data/yeastTF/) and on NCBI with
accession [GSE5499](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE5499). I used
the data provided by the author.    

Details on my parsing can be found in `scripts/`. The gene features are from
BrentLab/yeast_genome_resources.

This repo provides 3 datasets:

- **knockout**: Deletion/knockout perturbation normalized log2 fold changes.
- **metadata**: Transcription factor metadata including essentiality and QC status.
- **overexpression**: Overexpression perturbation normalized log2 fold changes.

## Usage

The python package `tfbpapi` provides an interface to this data which eases
examining the datasets, field definitions and other operations. You may also 
download the parquet datasets directly from hugging face by clicking on
"Files and Versions", or by using the huggingface_cli and duckdb directly.
In both cases, this provides a method of retrieving dataset and field definitions.

### `tfbpapi`

After [installing
tfbpapi](https://github.com/BrentLab/tfbpapi/?tab=readme-ov-file#installation), you can
adapt this [tutorial](https://brentlab.github.io/tfbpapi/tutorials/hfqueryapi_tutorial/)
in order to explore the contents of this repository.

### huggingface_cli/duckdb

You can retrieves and displays the file paths for each configuration of
the "BrentLab/hughes_2006" dataset from Hugging Face Hub.

```python
from huggingface_hub import ModelCard
from pprint import pprint

card = ModelCard.load("BrentLab/hughes_2006", repo_type="dataset")

# cast to dict
card_dict = card.data.to_dict()

# Get partition information
dataset_paths_dict = {d.get("config_name"): d.get("data_files")[0].get("path") for d in card_dict.get("configs")}

pprint(dataset_paths_dict)
```

If you wish to pull the entire repo, due to its size you may need to use an
[authentication token](https://huggingface.co/docs/hub/en/security-tokens).
If you do not have one, try omitting the token related code below and see if
it works. Else, create a token and provide it like so:

```python
from huggingface_hub import snapshot_download
import duckdb
import os

repo_id = "BrentLab/hughes_2006"

hf_token = os.getenv("HF_TOKEN")

# Download entire repo to local directory
repo_path = snapshot_download(
    repo_id=repo_id,
    repo_type="dataset",
    token=hf_token
)

print(f"\n✓ Repository downloaded to: {repo_path}")

# Construct path to the knockout parquet file
parquet_path = os.path.join(repo_path, "knockout.parquet")
print(f"✓ Parquet file at: {parquet_path}")
```

Use your favorite method of interacting with `parquet` files (eg duckDB, but you could
use dplyr in R or pandas, too).

```python
# Connect to DuckDB and query the parquet file
conn = duckdb.connect()

query = """
SELECT * 
FROM read_parquet(?)
WHERE regulator_locus_tag = 'CST6'
"""

result = conn.execute(query, [parquet_path]).fetchall()
print(f"Found {len(result)} rows for CST6")
```