Uploading data

This tutorial shows you how to upload a dataset for training a model.

What you need before getting started

Format your dataset as a 2-column CSV, including the full sequence of each variant and one or more columns for your measured properties.

Uploading your dataset

Run the command below, with your dataset’s path in brackets:

import pandas as pd
dataset = pd.read_csv("./example_dataset.csv")
dataset.head(2)

	sequence	Name	acetamide_normalized_fitness	isobutyramide_normalized_fitness	propionamide_normalized_fitness
0	WRHGDISSSNDTVGVAVVNYKMPRLHTAAEVLDNARKIAEMIVGMK...	Seq-1	NaN	-0.5174	NaN
1	WRHGDISSSNDTVGVAVVNYKMPRLHTAAEVLDNARKIAEMIVGMK...	seq-2	-2.1514	-0.5154	-1.1457

Using your data

To use the data with our suite of tools, create the dataset in the backend:

assay = session.data.create(dataset, "Dataset name", "Dataset description")
assay_id = assay.id
assay

AssayMetadata(assay_name='Dataset Name', assay_description='Dataset description', assay_id='b6ee60f0-05ad-4c55-ad25-8006b09220ba', original_filename='assay_data', created_date=datetime.datetime(2024, 5, 9, 5, 57, 9, 973856), num_rows=15, num_entries=41, measurement_names=['isobutyramide_normalized_fitness', 'acetamide_normalized_fitness', 'propionamide_normalized_fitness'], sequence_length=346)

Next steps

Our Data API page contains more information about using your dataset with the API.

You can use your uploaded dataset to train models, perform a single site analysis, and design sequences. For more information, see Training models, Single site analysis, and Designing sequences.