Wrapper around Dorado for basecalling Nanopore files
To run the Dorado basecaller, you need a machine with a GPU that supports CUDA drivers. According to public benchmarking data, g5.xlarge is the most cost-effective machine for basecalling. We use this EC2 instance with the Deep Learning Base OSS Nvidia Driver AMI (Amazon Linux 2) (AMI identifier: ami-069699babcf73a576).
- dorado 0.7.3: (https://github.com/nanoporetech/dorado)
- samtools: (https://www.htslib.org/download/)
- mount-s3: (https://github.com/awslabs/mountpoint-s3/tree/main)