class Preprocess(Task):
""" Unzip fastq files, keep only sequence info of those containing only ACGT """
mask = ['family', 'motif']
inputs = {'fastq': File('original', path=lambda x: "{}/{}.fastq.gz".format(download_dir, x['accession']), mask=mask)}
targets = {'seq': File('NoN.fastq', root=store_dir, mask=mask),
'cnt': File('NoN.fastq.cnt', root=store_dir, mask=mask)}
actions = [(unzip_seq_filter_N, ['#barcode', '$fastq', '$seq', '$cnt'])]
class GetCounts(Task):
"""Count lines, words and characters in file"""
inputs = {'inp': File('text', path=['a.txt', 'b.txt'])}
targets = {'out': File('counts.csv')}
actions = [
("(echo line word char file; wc {}) | sed 's/^ \+//;s/ \+/,/g' > {}",
["$inp", "$out"])
]
class CombineCoverage(Task):
mask = ['group', 'sample']
inputs = {'cov': GetCoverage.targets['cov']}
targets = {'csv': File('combined.csv', mask = mask),
'pdf': File('pltcov.pdf', mask = mask, root = '.')}
actions = [(combine_csvs, ['#cov', '#csv']),
("""echo "library(ggplot2); pdf('{}')
ggplot(read.csv('{}'), aes(x = val)) +
geom_density(aes(color = factor(sample)))"\
| R --vanilla""", ['$pdf','$csv'])]
class Partition(Task):
""" Partition aptamer sequences into motif-containing (fg) and motif-free (bg)
based on distance from MOTIF """
mask = ['family']
inputs = {'seq': Preprocess.targets['seq'],
#'nbr': File('nbr.txt', root=store_dir)
}
targets = {'fg': File('fg.txt', mask=mask, root=store_dir),
'bg': File('bg.txt', mask=mask, root=store_dir)}
actions = [(partition_aptamers, [fg_type, '$seq', '#motif', 'hahah', '$fg', '$bg', '#barcode'])]
class MergeAdaptivePrecisionRecall(Task):
inputs = {'inp': jf_roc_csv}
targets = {
'out': File('sim_precision_recall_multi_reg.tsv',
root=final_results_dir)
}
actions = [(combine_csvs, ['#inp', '#out', '\t'])]
class MergeConstrainedPrecisionRecall(Task):
inputs = {'inp': jf_const_roc_csv}
targets = {
'out': File('sim_precision_recall_single_reg.tsv',
root=final_results_dir)
}
actions = [(combine_csvs, ['#inp', '#out', '\t'])]
class GetBgSeqmers(Task):
""" Get seqmers in the motif-free pool"""
mask = ['family']
inputs = {'bg': Partition.targets['bg'],
'seq': Preprocess.targets['seq'],
'cnt': Preprocess.targets['cnt'],
}
targets = {'out': File('seqmer.txt', mask=mask, root=store_dir)}
actions = [(gen_bg_seqmers, ['$seq', '$cnt', '$bg', '$out', seqmer_len, '#barcode'])]
pdb_sim.add_param(sim_models, 'model')
pdb_sim.add_param(thetas, 'theta')
pdb_sim.add_param(sigmasqs, 'sigmasq')
pdb_sim.add_param(gammatildes, 'gammatilde')
pdb_sim.add_param(tree_sizes, 'tree')
pdb_sim.add_param(restrict_alphas)
pdb_sim.add_param(pd.read_csv(StringIO(reg_string), comment='#'))
pdb_sim.add_param(restrict_folds)
pdb_regime = ParamDb("regime")
pdb_regime.add_param(sim_regs, 'regime')
# Next define JUDI files associated with task: simulate
jf_reg = File('regime',
param=pdb_regime,
root=metadata_dir,
path=lambda x: 'regime_{}.csv'.format(x['regime']))
jf_modreg = jf_reg.copy().rename({'regime': 'modreg'})
jf_tree = File(
'newick',
param=pdb_tree,
root=metadata_dir,
path=lambda x: 'drosophila{}.newick'.format(48
if x['tree'] == 'big' else 9))
# ATTN.2 : We don't need to worry about the actual path of the files
# JUDI will create them automatically in a hierachry determined
# by the sorted order of parameters, e.g.,
# sim_data_dir/sim_data_tsv/alpha~1/fold~1/gammatilde~0.25/model~OU1/modreg~global/sigmasq~1/theta~1000/tree~small/sim_data.tsv
class AlignFastq(Task):
inputs = {'reads': File('orig_fastq', path = path_gen)}
targets = {'sai': File('aln.sai')}
actions = [('bwa aln {} {} > {}', [REF,'$reads','$sai'])]
class GetCoverage(Task):
mask = ['group']
inputs = {'bam': CreateBam.targets['bam']}
targets = {'cov': File('cov.csv', mask = mask)}
actions = [('(echo val; samtools rmdup {} - | samtools mpileup - | cut -f4) > {}', ['$bam','$cov'])]
class CreateBam(Task):
mask = ['group']
inputs = {'reads': AlignFastq.inputs['reads'],
'sai': AlignFastq.targets['sai']}
targets = {'bam': File('aln.bam', mask = mask)}
actions = [('bwa sampe {} {} {} | samtools view -Sbh - | samtools sort - > {}', [REF,'$sai','$reads','$bam'])]
class CountBgSeqmers(Task):
""" Count seqmers in the motif-free pool"""
mask = ['family']
inputs = {'inp': GetBgSeqmers.targets['out']}
targets = {'out': File('seqmer.txt.cnt', mask=mask, root=store_dir)}
actions = [('cat {} | sort | awk \'BEGIN {{OFS="\t"}} ($1 == last || last == "") {{sum += $2}} ($1 != last && last != "") {{print last, sum; sum = $2}} {{last = $1}} END {{print last, sum}}\' > {}', ['$inp', '$out'])]
class CombineCounts(Task):
"""Combine counts"""
mask = ['n']
inputs = {'inp': GetCounts.targets['out']}
targets = {'out': File('result.csv', mask=mask, root='.')}
actions = [(combine_csvs, ["#inp", "#out"])]
from judi import ParamDb, File
racer = ParamDb('racer')
racer.add_param(['tortoise', 'hare'], 'racer')
racer.add_param([1, 2], 'game')
game = ParamDb('game')
game.add_param([1, 2], 'game')
jf_racer = File('timing.csv', param=racer)
jf_game = File('combined_timing.csv', param=game)
class simulate(Task):
param = racer
targets = {'out': jf_racer}
actions = [('cp {}_{}.csv {}', ['#racer', '#game', '$out'])]
class combine(Task):
param = game
inputs = {'inp': jf_game}
targets = {'out': jf_racer}
actions = [(combine_csvs, ['#inp', '#out'])]