swhitfield/biomap-research-temperature_stability

temperature_stability

Sourced from biomap-research/temperature_stability and prepared for Hugging Face datasets usage.

Intended use

Prediction of protein stability across temperatures.

Provenance

• Card generated (UTC): 2026-10-06

• Upstream revision: unpinned (default branch; pin source_revision to a commit to reproduce the input).

• Upstream dataset card: https://huggingface.co/datasets/biomap-research/temperature_stability/blob/main/README.md (label definitions, license, citation).

Data files

Processed parquet files are stored under data/ using Hugging Face split naming conventions (train-*, validation-*, test-*).

Dataset statistics: stats.json at the dataset root. It includes row counts by split, columns, and SeqKit sequence-length metrics.

Preparation

• Preprocess mode: minimal.
• Seed: 1957723.
• No max sequence length filter was applied.
• Renamed source columns: label -> targets, seq -> sequence.
• Columns: id, sequence, targets, split.
• Validation handling: downloaded source split was preserved.
• Split subsets: canonical train/validation/test retain source assignments except any synthesized validation; alternate views are test_cluster, test_random, test_stratified, train_cluster, train_random, train_stratified, validation_cluster, validation_random, validation_stratified, generated with seed 1957723 from pooled canonical rows. They repartition the same 419235 canonical rows, not additional examples.

Sequence clustering

Configured settings for MMseqs2 cluster-based assignment, used when that path runs:

• MMseqs workflow: easy-linclust
• Minimum sequence identity (--min-seq-id): 0.3
• Minimum coverage (-c): 0.8
• Coverage mode (--cov-mode): 1
• Linclust version (--linclust-version): 2
• Threads: 1
• Seed (cluster ordering / tie-breaking): 1957723

Split sizes

• test: 73205 rows
• test_cluster: 73205 rows
• test_random: 73205 rows
• test_stratified: 73205 rows
• train: 283057 rows
• train_cluster: 283057 rows
• train_random: 283057 rows
• train_stratified: 283057 rows
• validation: 62973 rows
• validation_cluster: 62973 rows
• validation_random: 62973 rows
• validation_stratified: 62973 rows

Loading

from datasets import load_dataset

ds = load_dataset("swhitfield/biomap-research-temperature_stability")
print(ds)