Borzoi Chromatin Accessibility MORSE

Score a DNA target for a Morse-code chromatin accessibility pattern using Borzoi.

License: Borzoi uses Apache-2.0 for code and CC-BY-4.0 for model weights and may require explicit attribution when utilized. Please refer to the code license and model weights license for full terms.

This constraint is open source. Any third-party models, product names, or trademarks referenced are the property of their respective owners, and Proto is not affiliated with them.

---

Tools Used Tools Used Source Source Cite Cite

Go to Tool Page

proto-bio/proto-language/proto_language/constraint/sequence_annotation/borzoi_chromatin_accessibility_morse_constraint.py

View source

@article{linder2025borzoi,
  title={Predicting RNA-seq coverage from DNA sequence as a unifying model of gene regulation},
  author={Linder, Johannes and Srivastava, Divyanshi and Yuan, Han and Agarwal, Vikram and Kelley, David R},
  journal={Nature Genetics},
  volume={57},
  number={4},
  pages={949--961},
  year={2025},
  publisher={Nature Publishing Group},
  doi={10.1038/s41588-024-02053-6}
}

Copy citation

Score target proposals against a Borzoi chromatin accessibility MORSE objective.

API Reference

ConfigBorzoiChromatinAccessibilityMorseConfig Source

Configuration for Borzoi chromatin accessibility MORSE scoring.

organism · enum · default: "human"

Borzoi species head to use: human or mouse.

Options: human, mouse

pattern · string · default: ".--. .-. --- - ---"

Morse pattern; use ’.’, ’-’, and spaces between letters.

dot_bp · integer · default: "25"

Length of each dot window.

dash_bp · integer · default: "75"

Length of each dash window.

intra_symbol_gap_bp · integer · default: "25"

Gap between symbols within a letter.

inter_letter_gap_bp · integer · default: "75"

Gap between letters in the Morse pattern.

pattern_start_bp · integer · default: "0"

Target-relative start coordinate where the Morse pattern begins.

pattern_normalization · enum · default: "global_max"

Scale signal by full-output max, target-region max, or not at all.

Options: global_max, region_max, none

contrast_margin · number · default: "0.0"

Minimum normalized high-minus-low signal margin to reward.

contrast_weight · number · default: "0.0"

Penalty weight when the normalized contrast margin is too small.

raw_amplitude_weight · number · default: "0.0"

Reward weight for larger raw target signal range.

high_window_reward_weight · number · default: "0.0"

Reward weight for larger raw signal in dot and dash windows.

low_window_penalty_weight · number · default: "0.0"

Penalty weight for larger raw signal in gap windows.

window_stat_transform · enum · default: "log1p"

Transform for raw amplitude and window-mean reward terms.

Options: log1p, identity

device · string · default: "cuda"

CUDA device for Borzoi inference.

batch_size · integer · default: "1"

Candidate sequences per Borzoi model batch.

trim_prefix_bp · integer · default: "0"

Leading target bases to ignore before accessibility scoring.

borzoi_output_tracks · List[integer] · default: "[1901]"

Borzoi chromatin-accessibility tracks; defaults by organism.

borzoi_ensemble_reduce_method · enum · default: "mean"

How to combine Borzoi replicate signals.

Options: mean, min, std, lcb

Usage

python

from proto_language.core import Constraint
from proto_language.constraint import borzoi_chromatin_accessibility_morse_constraint, BorzoiChromatinAccessibilityMorseConfig

constraint = Constraint(
    inputs=[segment],
    function=borzoi_chromatin_accessibility_morse_constraint,
    function_config=BorzoiChromatinAccessibilityMorseConfig(
        # Configure parameters here
    ),
)

scores = constraint.evaluate()

Metadata

Property	Value
Key	borzoi-chromatin-accessibility-morse
Function	borzoi_chromatin_accessibility_morse_constraint
Category	sequence_annotation
Mode	discrete
Uses GPU	True
Supported Types	dna