gpn

GPN (Genomic Pre-trained Network)

Code and resources from GPN paper and GPN-MSA paper.

Table of contents

• Installation
• Minimal usage
• GPN
• GPN-MSA
• Citation

Installation

pip install git+https://github.com/songlab-cal/gpn.git

Minimal usage

import gpn.model
from transformers import AutoModelForMaskedLM

model = AutoModelForMaskedLM.from_pretrained("songlab/gpn-brassicales")
# or
model = AutoModelForMaskedLM.from_pretrained("songlab/gpn-msa-sapiens")

GPN

Can also be called GPN-SS (single sequence).

Examples

• Play with the model: examples/ss/basic_example.ipynb

Code and resources from specific papers

• Arabidopsis thaliana

Training on your own data

1. Snakemake workflow to create a dataset
   • Can automatically download data from NCBI given a list of accessions, or use your own fasta files.
2. Training
   • Will automatically detect all available GPUs.
   • Track metrics on Weights & Biases
   • Implemented encoders: convnet (default), roformer (Transformer), bytenet
   • Specify config overrides: e.g. --config_overrides encoder=bytenet,num_hidden_layers=30
   • The number of steps that you can train without overfitting will be a function of the size and diversity of your dataset
   • Example:

WANDB_PROJECT=your_project torchrun --nproc_per_node=$(echo $CUDA_VISIBLE_DEVICES | awk -F',' '{print NF}') -m gpn.ss.run_mlm --do_train --do_eval 
    --report_to wandb --prediction_loss_only True --remove_unused_columns False 
    --dataset_name results/dataset --tokenizer_name gonzalobenegas/tokenizer-dna-mlm 
    --soft_masked_loss_weight_train 0.1 --soft_masked_loss_weight_evaluation 0.0 
    --weight_decay 0.01 --optim adamw_torch 
    --dataloader_num_workers 16 --seed 42 
    --save_strategy steps --save_steps 10000 --evaluation_strategy steps 
    --eval_steps 10000 --logging_steps 10000 --max_steps 120000 --warmup_steps 1000 
    --learning_rate 1e-3 --lr_scheduler_type constant_with_warmup 
    --run_name your_run --output_dir your_output_dir --model_type GPN 
    --per_device_train_batch_size 512 --per_device_eval_batch_size 512 --gradient_accumulation_steps 1 --total_batch_size 2048  
    --torch_compile 
    --ddp_find_unused_parameters False 
    --bf16 --bf16_full_eval 

3. Extract embeddings
   • Input file requires chrom, start, end
   • Example:

torchrun --nproc_per_node=$(echo $CUDA_VISIBLE_DEVICES | awk -F',' '{print NF}') -m gpn.ss.get_embeddings windows.parquet genome.fa.gz 100 your_output_dir 
    results.parquet --per_device_batch_size 4000 --is_file --dataloader_num_workers 16

4. Variant effect prediction
   • Input file requires chrom, pos, ref, alt
   • Example:

torchrun --nproc_per_node=$(echo $CUDA_VISIBLE_DEVICES | awk -F',' '{print NF}') -m gpn.ss.run_vep variants.parquet genome.fa.gz 512 your_output_dir results.parquet 
    --per_device_batch-size 4000 --is_file --dataloader_num_workers 16

GPN-MSA

Examples

• Play with the model: examples/msa/basic_example.ipynb
• Variant effect prediction: examples/msa/vep.ipynb
• Training (human): examples/msa/training.ipynb

Code and resources from specific papers

• Human

Training on other species (e.g. plants)

• See #28
• Another source for plant alignments: https://plantregmap.gao-lab.org/download.php#alignment-conservation

Citation

GPN:

@article{benegas2023dna,
    author = {Gonzalo Benegas  and Sanjit Singh Batra  and Yun S. Song },
    title = {DNA language models are powerful predictors of genome-wide variant effects},
    journal = {Proceedings of the National Academy of Sciences},
    volume = {120},
    number = {44},
    pages = {e2311219120},
    year = {2023},
    doi = {10.1073/pnas.2311219120},
    URL = {https://www.pnas.org/doi/abs/10.1073/pnas.2311219120},
    eprint = {https://www.pnas.org/doi/pdf/10.1073/pnas.2311219120},
}

GPN-MSA:

@article{benegas2023gpnmsa,
	author = {Gonzalo Benegas and Carlos Albors and Alan J. Aw and Chengzhong Ye and Yun S. Song},
	title = {GPN-MSA: an alignment-based DNA language model for genome-wide variant effect prediction},
	elocation-id = {2023.10.10.561776},
	year = {2023},
	doi = {10.1101/2023.10.10.561776},
	publisher = {Cold Spring Harbor Laboratory},
	URL = {https://www.biorxiv.org/content/early/2023/10/11/2023.10.10.561776},
	eprint = {https://www.biorxiv.org/content/early/2023/10/11/2023.10.10.561776.full.pdf},
	journal = {bioRxiv}
}