intial version

This commit is contained in:
Tom Kasper
2026-09-17 22:16:42 +01:00
parent e23ed24fa9
commit 08467102c8
19 changed files with 311 additions and 0 deletions
+68
View File
@@ -0,0 +1,68 @@
def get_model_name(wildcards):
config_param=f'{wildcards.model}_model_name'
return config[config_param]
def get_model_requirement(wildcards):
return f"{config['dorado_model_dir']}/{config[f'{wildcards.model}_model_name']}
def get_benchmarking_file(wildcards):
config_param = f'{wildcards.model}_benchmarks_file'
if config[config_param]:
return f'--batchsize-benchmarks-file {config[config_param]}'
return ''
def get_qscore(wildcards):
config_param = f'{wildcards.model}_min_q'
if config[config_param]:
return f'--min-qscore {config[config_param]}'
return ''
def get_batch_names(input_file):
with open(input_file,'rt') as ih:
return [line.strip().split('/')[-1].split('.')[0] for line in ih]
rule download_pod5:
output:
'../data/raw_pod5/PBK98658_853a956f_57f83f46_{batch}.pod5'
threads: 1
wildcard_constraints:
batch="\d+"
shell:
"""
aws s3 cp --no-sign-request s3://ont-open-data/nomiss_96BC_P2I_SUP_2026/raw/pod5/PBK98658_853a956f_57f83f46_{wildcards.batch}.pod5 {output}
"""
rule basecall_pod5:
input:
pod5='../data/raw_pod5/PBK98658_853a956f_57f83f46_{batch}.pod5',
model=get_model_requirement
output:
'../data/basecalled_reads/{model}/PBK98658_853a956f_57f83f46_{batch}.fastq.gz'
threads:
32
resources:
gpu=1
params:
benchmarking=get_benchmarking_file,
min_qscore=get_qscore,
dorado_model=get_model_name
wildcard_constraints:
batch="\d+",
model="hac|fast"
shell:
"""
dorado basecaller --models-directory {config[dorado_model_dir]} --emit-fastq {params.benchmarking} {params.min_qscore} {params.dorado_model} {input.pod5} > {output}
"""
rule concatenate_basecalled_fastq:
input:
expand('../data/basecalled_reads/{{model}}/PBK98658_853a956f_57f83f46_{batch}.fastq.gz',batch=range(1,config["pod5_dataset_size"]+1,config["pod5_stride"]))
output:
'../data/basecalled_reads/{model}.fastq.gz'
threads: 1
wildcard_constraints:
model="hac|fast"
shell:
"""
zcat {input} > {output}
"""