Update README.md

README.md

@@ -254,4 +254,113 @@ configs:
       - split: train
         path:
           - data/*/*/part-0.parquet
----
+---
+# Harbison 2004
+
+This Dataset is a parsed version of the data provided by Richard A. Young's lab. To cite this data, please use:  
+
+[Harbison CT, Gordon DB, Lee TI, Rinaldi NJ, Macisaac KD, Danford TW, Hannett NM, Tagne JB, Reynolds DB, Yoo J, et al. 2004. Transcriptional regulatory code of a eukaryotic genome. Nature 431: 99–104.doi:10.1038/nature02800](https://www.nature.com/articles/nature02800)
+
+## Usage
+
+You may access just the Dataset metadata like this:
+
+```python
+from huggingface_hub import ModelCard
+
+card = ModelCard.load("BrentLab/harbison_2004", repo_type="dataset")
+
+# cast to dict
+card_dict = card.data.to_dict()
+
+# Get partition information
+card_dict.get("dataset_info").get("partitioning").get("keys")
+```
+
+Output:
+
+```raw
+[{'name': 'condition',
+  'dtype': 'string',
+  'levels': [
+   'YPD',
+   'SM',
+   'RAPA',
+   'H2O2Hi',
+   'H2O2Lo',
+   'Acid',
+   'Alpha',
+   'BUT14',
+   'BUT90',
+   'Thi-',
+   'GAL',
+   'HEAT',
+   'Pi-',
+   'RAFF']},
+ {'name': 'regulator_symbol',
+  'dtype': 'string',
+  'levels': [
+   'YSC0017',
+   'YKL112W',
+   'YNR054C',
+   'YER045C',
+   ... (there are 204 regulators)
+```
+
+You can use this information to pull only the partition you're interested in, eg
+
+```python
+from huggingface_hub import hf_hub_download
+
+# Download a single partition file and store it in the local cache
+local_file = hf_hub_download(
+    repo_id="BrentLab/harbison_2004",
+    repo_type="dataset",
+    filename="data/condition=YPD/regulator_locus_tag=YER045C/part-0.parquet"
+)
+```
+
+<details>
+<summary><strong>Dataset Details</strong></summary>
+
+The data was extract from The [Young Lab's website](http://younglab.wi.mit.edu/regulatory_code/GWLD.html).
+
+I pulled the 'Excel' versions of both the P values (the downloaded file is called files_for_paper_abbr.zip)
+and Binding ratios (Ratio_forpaper_abbr.zip).  
+
+The following is copied from the [Young's lab explanation of the analysis](http://younglab.wi.mit.edu/regulatory_code/Analysis.html):  
+
+> The microarrays were scanned using an Axon200B scanner, and the images were analyzed with Genepix 5.0. Columns corresponding to the background
+> subtracted intensities and standard deviation of the background were extracted for further analysis. The intensities for the two channels, representing
+> the immunoprecipitated (test) and unenriched (control) samples, were normalized by using the median of each channel to calculate a normalization factor,
+> normalizing all datasets to a single median intensity. The log ratio of the intensity in the test channel to the control channel was calculated.
+> To account for biases in the immunoprecipitation reaction, these log ratios were normalized for each spot by subtracting the average log ratio of each
+> spot across all arrays. The intensities in the test channel were then adjusted to yield this normalized ratio. Finally, an error model was used to
+> calculate significance of enrichment on each chip and to combine data for replicates to obtain a final average ratio and significance of enrichment
+> for each intergenic region. Each intergenic region was assigned to the genes it is most likely to regulate.
+> We have included new refinements in our analysis relative to that used in Lee et al. Notably, we have excluded artefactual spots from analysis,
+> selected more reliable probes for normalization and assigned quality metrics to individual arrays to identify low quality experiments.
+
+The script used to parse this data from the data provided by the Young lab into the parquet dataset presented here is included in `scripts/`
+
+</details>
+
+<details>
+<summary><strong>Dataset Structure</strong></summary>
+
+### data/
+
+This is a **Parquet** dataset where the partitions are based on `condition` and `regulator_locus_tag`.
+Each row represents the effect and pvalue on a given target gene (`target_locus_tag`)
+
+| Field                 | Description                                                                                                                                                                                      |
+|-----------------------|-------------------------------------------------------------------------------------------------------------------------|
+| `condition`           | See below for a definition of each level                                                                                |
+| `regulator_locus_tag` | The systematic ID of the ChIP'ed regulator. See hf/BrentLab/yeast_genome_resources to map to the common name (symbol)   |
+| `target_locus_tag`    | the systematic ID of the feature to which the Young lab assigned the effect/pvalue                                      |
+| `effect`              | chip based binding ratio                                                                                                |
+| `pvalue`              | pvalue of the effect                                                                                                    |
+
+</details>
+
+**Dataset Author and Contact**: Chase Mateusiak [@cmatKhan](https://github.com/cmatkhan/)