NEMO: a NEural network model for mapping MOdifications in nanopore Long-read

🚀 Overview

NEMO is a deep learning tool designed to predict DNA modifications using nanopore long-read chromatin accessibility data. It allows users to train neural network models, predict modifications, and visualize results.

⚡ Installation

1. Clone the repository

git clone https://github.com/baigal628/NEMO.git
cd NEMO

2. Create and activate the conda environment:

conda create -n nemo python=3.9
conda activate nemo

3. Install dependencies:

pip install -r requirements.txt

🛠️ Basic utilities

Navigate to Test Scripts

cd NEMO/test/sh/

🔄 Data preprocessing for nanopore sequencing data

Basecall data using dorado: https://github.com/nanoporetech/dorado

dorado basecaller [email protected] \
    ../input/test.pod5 \
    --emit-moves \
    --device cuda:all \
    --reference ../input/sacCer3.fa > ../input/test_reads.bam

Signal-to-Event Alignment: https://github.com/cafelton/pod5-to-kmer-signal

python3 ../../src/ref/bampod5kmersig-witharrow-sigalign.py \
    -b ../input/test_reads.bam \
    -p ../input/test.pod5
    -o ../input/test

📈 Train and test model using positive and negative control data

Preprocessed negative and positive data ar provided under ./nemo/test/output/

python3 ../../src/train.py \
    --exp_id test_r10 \
    --neg_data ../output/can-sigalign.parquet \
    --pos_data ../output/mod-sigalign.parquet \
    --batch_size 256 \
    --seq_len 400 \
    --model_type resnet \
    --outpath ../output/ \
    --save_test \
    --epochs 5 \
    --steps_per_epoch 20 \
    --val_steps_per_epoch 10 

python3 ../../src/test.py \
    --exp_id test_r10 \
    --model_type resnet \
    --test_dataset ../output/test_dataset_test_r10_resnet.pt \
    --weight ../output/test_r10_resnet_best_model.pt \
    --outpath  ../output/ \
    --batch_size 256

🔍 Predict modifications on chromatin data

python3  ../../src/predict.py \
    --bam ../input/test_reads.bam \
    --parquet ../output/test.parquet \
    --region 'chrI:500-2500' \
    --seq_len 400 \
    --step 200 \
    --weight ../output/test_r10_resnet_best_model.pt \
    --thread 4 \
    --outpath ../output/ \
    --prefix mod_prediction \
    --batch_size 216 \

📊 plot metagene at TSS

python3  ../../src/plot.py \
    --plot aggregate \
    --pred ../output/mod_prediction.tsv \
    --bed TSS.bed \
    --ref ../input/sacCer3.fa \
    --label 6mA \
    --outpath ../output/ \
    --prefix mod_prediction

Reference

Probing chromatin accessibility with small molecule DNA intercalation and nanopore sequencing

Gali Bai*, Namrita Dhillon*, Colette Felton*, Brett Meissner*, Brandon Saint-John*, Robert Shelansky*, Elliot Meyerson, Eva Hrabeta-Robinson, Babak Hodjat, Hinrich Boeger, Angela N. Brooks bioRxiv 2024.03.20.585815; doi: https://doi.org/10.1101/2024.03.20.585815

📬 Feedback & Contributions

We welcome contributions! Feel free to submit issues or pull requests to improve NEMO.

✨ Acknowledgments

Developed with ❤️ by Brooks Lab and Cognizant AI Labs. Thanks to the contributors and open-source community for their support!