pyeast 2.0.0

A Python toolkit for yeast engineering

PYEAST

PYEAST is a command-line toolkit and Python library that automates the design of DNA cloning experiments in Saccharomyces cerevisiae (baker's yeast). Given a set of genetic parts, it designs the PCR primers and liquid-handling instructions for common yeast genetic engineering techniques — reducing manual design work and minimising errors.

If you are new to yeast cloning, PYEAST is designed to complement standard wet-lab protocols: it handles the computational design steps so you can focus on the biology.

For full methodological details, see our publication.

Prerequisites

• Python (3.12 or later) — python.org
• Git — git-scm.com
• uv - docs.astral.sh

Installation

uv tool install git+https://github.com/TomLoan/PYEAST.git

# Download the sequence data automatically
pyeast init

pyeast init clones the data repository to ~/PYEAST/data/ and configures PYEAST automatically. If you have already downloaded the data elsewhere, point PYEAST at it instead:

pyeast init --data-dir /path/to/PYEAST_data

Commands

Run pyeast --help to see all available commands:

Command	Description
pyeast tar	Design primers for Transformation-Assisted Recombination (TAR) cloning
pyeast integrate	Design primers for chromosomal integration
pyeast delete	Design deletion cassettes (scarless marker-recycling method)
pyeast replace	Design replacement cassettes (scarless marker-recycling method)
pyeast gg	Design Golden Gate / MoClo assemblies
pyeast batch	Regenerate instruction files for previously designed assemblies
pyeast agent	Interactive LLM-powered experiment design session (default backend needs an extra)
pyeast init	Configure the data directory

For help with any command: pyeast COMMAND --help

TAR Cloning

Transformation-Assisted Recombination (TAR) exploits S. cerevisiae's natural homologous recombination to assemble PCR products into new plasmids in a single yeast transformation. PYEAST designs all primers, including the homology overhangs, for each part in your assembly. PYEAST uses pydna to model the PCR and homologous recombination steps and writes a cloning-history file (_history.json) that you load history in OpenCloning to visualise and share how the plasmid was assembled.

Chromosomal Integration

Design primers for inserting a DNA construct into a specific locus in the yeast genome. PYEAST calculates the flanking homology sequences required for efficient integration. As with TAR cloning, the integration is modelled with pydna and exported as an OpenCloning history file for visualising the assembly.

Gene Deletion and Replacement

Design cassettes for deleting or replacing genes using the scarless marker-recycling method described by Akada et al., 2006.

Beta feature: Golden Gate / MoClo Assembly

== Beta feature, please double check outputs ==

Design Golden Gate assemblies using the MoClo standard. Uses DNA Cauldron to simulate the assembly. The following part libraries are included out of the box:

Kit	Reference
Yeast ToolKit (YTK)	Lee et al., 2015 — Addgene
Yeast Secretion and Display Toolkit	O'Riordan et al., 2023 — Addgene
OPENPichia MoClo Kit	Claes et al., 2024 — BCCM

To add parts to an existing kit, save a FASTA file to component_libraries/<kit_name>/ in your data directory. For liquid-handling support, also save the matching plasmid .gb file to component_libraries/<kit_name>/plasmids/ and add a well entry to templates/TemPlates.xlsx.

To add a new Golden Gate kit, create a new subfolder in component_libraries/ and follow the same structure.

Beta feature: LLM Agent

== Beta feature, please double check outputs ==

pyeast agent starts an interactive session where you describe your experiment in plain language. The agent discovers available components, looks up gene sequences from SGD, designs TAR cloning, chromosomal integrations, gene deletions and replacements, and generates consolidated PCR batch instructions — calling the same underlying functions as the CLI commands.

Note: The agent is an interface layer over the standard design tools. Double-check outputs against direct CLI command results if you have any doubts.

Installation

The default Anthropic backend requires an extra that is not installed by default. Enable it by installing PYEAST with the agent extra:

uv tool install "git+https://github.com/TomLoan/PYEAST.git#egg=pyeast[agent]"

If you have already installed PYEAST, re-run the command above to add the extra. Running the default backend without it exits with an error telling you to install pyeast[agent].

The Ollama and OpenAI-compatible backends (below) need no extra dependency and work with a standard install. All other PYEAST commands work without the extra too.

Usage

Three LLM backends are supported. Copy .env.example to .env and fill in the relevant keys.

Anthropic (default) — requires ANTHROPIC_API_KEY.

pyeast agent

Ollama — requires a locally running Ollama server with a tool-calling model.

ollama serve && ollama pull qwen2.5
pyeast agent --provider ollama --model qwen2.5

OpenAI-compatible servers (e.g. LM Studio) — requires a tool-calling model loaded and the server started in LM Studio's Developer tab.

pyeast agent --provider openai --model <model-key>
# or, if your server is not on the default port (1234):
pyeast agent --provider openai --model <model-key> --base-url http://localhost:5678/v1

Private Data

You can store proprietary sequences, primers, and templates in a private/ subdirectory of your data folder. PYEAST searches both public and private locations automatically, and the private folder is excluded from version control.

<data directory>/
├── component_libraries/       ← public parts
├── integration_sites/
├── primers/
├── templates/
└── private/
    ├── component_libraries/   ← your private parts (gitignored)
    ├── integration_sites/
    ├── primers/
    └── templates/

When you run pyeast init, the private/ folder structure is created automatically.

Configuration

PYEAST looks for data in this order:

1. Environment variable: PYEAST_DATA_DIR
2. Config file: ~/PYEAST/config.yaml
3. Default: ~/PYEAST/data/

To view or update your current configuration, run pyeast init.

Troubleshooting

"Data directory not found"

Run pyeast init to check what path is configured. To reconfigure:

pyeast init --data-dir /correct/path/to/PYEAST_data

Commands fail with missing files

Verify that your data directory contains the expected folder structure and that FASTA/Excel files are present in the right locations.

Import errors on startup

Reinstall to ensure all dependencies are present:

pip install git+https://github.com/TomLoan/PYEAST.git

Python API

All designer classes can be imported and used directly — no interactive prompts, no CLI required. Useful for scripting, Jupyter notebooks, or integrating PYEAST into larger pipelines.

Each designer follows the same pattern: create a designer, call design() with your inputs, get back a result object, and call result.save() to write files.

from pathlib import Path
from pyeast.core.tar import TARDesigner
from pyeast.utils.path_utils import get_component_libraries_path, get_primers_path, get_templates_path

designer = TARDesigner(homology_length=25)
result = designer.design(
    library_path=get_component_libraries_path() / "YeastToolKit",
    assembly_order=["pTEF1", "YeGFP", "tCYC1", "Ura3", "AmpR_ColE1", "2Micron"],
    primer_folder=get_primers_path(),
    template_folder=get_templates_path(),
    name="my_plasmid",
)
result.save(Path("output/my_plasmid/my_plasmid"))

The same designer.design() → result.save() pattern applies to IntegrationDesigner, DeletionDesigner, and ReplaceDesigner. See docs/ or run help(ClassName) for constructor options and full parameter details.

For Developers

git clone https://github.com/TomLoan/PYEAST.git
cd PYEAST
uv sync --group dev
uv run pytest tests/ -v

Contributions are welcome — see CONTRIBUTING.md for details.

Citation

If you use PYEAST in your research, please cite:

Madika, A., Suri, A., Purohit, A. et al. PYEAST – A Computational Toolkit for Saccharomyces cerevisiae Genetic Engineering. npj Syst Biol Appl 12, 83 (2026). https://doi.org/10.1038/s41540-026-00712-4