Nanopore sequencing is a third generation approach used in the sequencing of biopolymers-specifically, polynucleotides in the form of DNA or RNA.
GeneSeq_BaseCaller is developed to build AI applications for Gene Sequence Base calling.
# build docker image
docker image build -t lt_basecaller:0.1 . -f Dockerfile
# create container and run
docker run --gpus all -p 8899:8899 --name basecaller_dev --ipc=host -it -v /workspace/GeneSeq_BaseCaller:/workspace/GeneSeq_BaseCaller lt_basecaller:0.1
cd GeneSeq_BaseCaller
# setup
python setup.py develop# to download ont data
python bonito/download.py --training --data_dir /workspace/datasets/
# to download ont models
python bonito/download.py --model --show
python bonito/download.py --model --model_dir /workspace/model_bin/If you have your own hdf5 file, then convert it to numpy data directly.
python bonito/convert.py /workspace/datasets/*.fast5 /workspace/datasets/bash scripts/train.sh
bash scripts/val.shpython bonito/basecaller.py /workspace/model_bin/MODEL /workspace/datasets/READs -v --reference /workspace/datasets/REF > basecalls.bam| dataset | model | mean_acc | device | epoch |
|---|---|---|---|---|
| dna_r9.4.1 | lstm_ctc_crf | 96.37 | A10 | 20 |
| dna_r9.4.1 | quartznet | 95.72 | A10 | 20 |
| dna_r9.4.1 | TransEnc_ctc_crf | 96.04 | A10 | 10 |
| dna_r9.4.1 | conformer_ctc_crf | 97.18 | A100 | 10 |
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent