Skip to content

Ecosystem Overview

The TSL Structures project provides two main tools for working with protein structure datasets.

Tools

tsp-maker (Python)

Purpose: Create TSP packages from prediction outputs

Install: 

pip install git+https://github.com/TeamMacLean/tsp-maker.git

Capabilities: - Parse outputs from AlphaFold2, AlphaFold3, or Boltz2 - Build TSP packages with standardised metadata - Upload datasets to Zenodo

Documentation: teammaclean.github.io/tsp-maker

tslstructures (R)

Purpose: Find, download, and analyse TSP datasets

Install: 

# install.packages("pak")
pak::pak("TeamMacLean/tslstructures")

Capabilities: - Discover datasets in the TSL Structures Zenodo community - Download and cache datasets locally - Query metadata and filter structures - Extract structures for analysis

Documentation: teammaclean.github.io/tslstructures

How They Work Together

                                    ┌─────────────────────┐
                                    │  Prediction Tools   │
                                    │  (AF2, AF3, Boltz2) │
                                    └──────────┬──────────┘
                                               │
                                               ▼
┌──────────────────────────────────────────────────────────────────────┐
│                           DATA PRODUCER                               │
│                                                                       │
│  1. Run predictions                                                   │
│  2. tsp-maker parse af3 /predictions /intermediate                   │
│  3. tsp-maker build /intermediate /my-dataset --name my-structures   │
│  4. tsp-maker upload /my-dataset --publish                           │
│                                                                       │
└──────────────────────────────────────────────────────────────────────┘
                                               │
                                               ▼
                                    ┌─────────────────────┐
                                    │       Zenodo        │
                                    │   (with DOI)        │
                                    └──────────┬──────────┘
                                               │
                                               ▼
┌──────────────────────────────────────────────────────────────────────┐
│                           DATA CONSUMER                               │
│                                                                       │
│  datasets <- list_datasets()                                          │
│  install_dataset("10.5281/zenodo.12345678")                          │
│  ds <- load_dataset("my-structures")                                  │
│  high_conf <- ds |> filter(plddt_mean > 90)                          │
│  structure <- get_structure(ds, "P12345_AF3_1")                      │
│                                                                       │
└──────────────────────────────────────────────────────────────────────┘

Component Summary

Component Language Purpose For
tsp-maker Python Create and upload datasets Producers
tslstructures R Find and use datasets Consumers
Zenodo - Host datasets with DOIs Distribution

Further Reading