prism-bio 1.0.3

PCR primer design & optimization pipeline

PRISM - Primer Design through Submodular Function Estimation

User Guide

PRISM (PRImer Selection through Submodular Maximization) is an open-source software tool for automated design of multiplex PCR primers, developed to support viral genome sequencing, pathogen surveillance, and other high-throughput molecular biology applications. Released under the GNU General Public License (GPL), PRISM is freely available and encourages community use, extension, and collaboration.

PRISM introduces a principled optimization framework that formulates the primer design task as a constrained submodular maximization problem, balancing two competing objectives: maximizing genome coverage and minimizing undesired primer-primer interactions, quantified using the Badness value. This formulation allows PRISM to leverage a fast local search algorithm with a provable constant-factor approximation guarantee, making it the first primer design method to combine rigorous theoretical guarantees with practical scalability.

In extensive evaluations on viral genome datasets such as foot-and-mouth disease virus (FMDV) and Zika virus, PRISM consistently outperforms leading tools—including PrimalScheme, Olivar, and primerJinn—achieving significantly lower Badness values, tighter distributions of primer quality, and robust genome coverage with low memory usage and runtime.

---

Installation

Requirements

– Python ≥ 3.9.
– Linux, macOS or Windows

Runtime dependencies (primer3‑py, numpy, pandas, tqdm, numba, joblib) are installed automatically.

Installation from PyPI using pip

Recommend create and activate a virtual environment:

python3 -m venv /path/to/prism
source /path/to/prism/bin/activate

Or using conda:

conda create -n prism python=3.9 -y
conda activate prism

Please use the following command to install：

pip install prism-bio

Once the installation is complete, use the following commands to check PRISM's version:

prism --version

---

General usage

Input files

PRISM requires a reference genome in FASTA format as input. This file should contain one or more nucleotide sequences in standard FASTA format. For optimal results in multiplex primer design, the input should represent a high-quality reference or consensus sequence derived from your target population or viral strain.

---

Command-line interface

PRISM accepts three main sub-commands, input (required), windows (optional, with a default size of 250 bp), and output (optional, specifies file name and output path):

CLI options

Flag	Default	Description
-i, --input	required	Input FASTA/FA file
-w, --window-size	250	Window size for region slicing
-o, --output-csv	optimized_primers.csv	Output file (CSV)

Full help:

prism --help
or
prism -h

---

Example usage

# Run PRISM on a Zika reference, 300 bp windows, export CSV
prism \
  -i data/NC_012532.1.fna \
  -w 250 \
  -o output/optimized_primers.csv

You can find this example file NC_012532.1.fna in example input. This is the Zika virus sequence from the NCBI dataset.

---

Generate files

The process ends with a .csv file in the form of the following table:

Primer ID	Left Primer	Right Primer
Primer 1	GTGTGA…	CGTAGC…
Primer 2	GCGTAC…	TAGCCA…
Primer…	………………	………………

This file will provide the designed primer serial numbers as well as the sequence

---

Developer Setup

git clone https://github.com/William-A-Wang/PRISM.git
python -m venv .venv && . .venv/Scripts/activate
pip install -e .[dev]

Using this approach, you can deploy code locally while performing development-related activities

---

External resources

We used Olivar's badness calculation function in our code, but modified some of the parameters.

• Olivar

---

License

PRISM is released under the GNU GPL v3.0.

See the LICENSE file for details.

---

Citing PRISM

If you use PRISM in an academic setting, please cite:

@misc{WangPRISM25,
author      = { Ao Wang and 
                Yixin Chen and
                Aaron Hong and
                Adam Rivers and
                Alan Kuhnle and
                Christina Boucher},
title       = {Primer Design through Submodular Function Estimation},
note        = {In submission}
year        = {2025},
}