Unverified Commit 049f57c1 authored by Chris Cheshire's avatar Chris Cheshire Committed by GitHub
Browse files

Merge pull request #21 from luslab/dev 

Included workflow solution for different experimental-igg replicate number combinations
parents 8a79b87b edf3a2ea

bin/lib/reports.py

+9 −3
Original line number Diff line number Diff line
@@ -277,6 +277,8 @@ class Reports: 
        unique_replicates = self.seacr_beds.replicate.unique()

        rep_permutations = array_permutate(range(len(unique_replicates)))

        self.replicate_number = len(unique_replicates)

        # print("this is rep_permutations")

        # print(rep_permutations)



        if self.replicate_number > 1:

            idx_count=0
@@ -297,9 +299,13 @@ class Reports: 
                        else:

                            pyr_query = pyr_subject



                    pyr_starts = pyr_overlap.values()[0]['Start']

                    if (pyr_query.empty):

                        self.reprod_peak_stats.at[idx_count, 'no_peaks_reproduced'] = 0

                    else :

                        pyr_starts = pyr_query.values()[0]['Start']

                        unique_pyr_starts = pyr_starts.unique()

                        self.reprod_peak_stats.at[idx_count, 'no_peaks_reproduced'] = len(unique_pyr_starts)



                    idx_count = idx_count + 1



            fill_reprod_rate = (self.reprod_peak_stats['no_peaks_reproduced'] / self.reprod_peak_stats['all_peaks'])*100

conf/igg_flow_tests/test_all_exp_one_igg_small.config

0 → 100644
+23 −0
Original line number Diff line number Diff line 
params {

    config_profile_name        = 'Test profile'

    config_profile_description = 'Minimal test dataset to check pipeline function'



    // Limit resources so that this can run CI

    max_cpus   = 2

    max_memory = 6.GB

    max_time   = 6.h



    // Input data

    input = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/samplesheet/test-GSE145187-allexp-1igg-small.csv'



    // Genome references

    fasta              = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20.fa.gz'

    bowtie2            = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-bowtie2.tar.gz'

    gtf                = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.gtf.gz'

    gene_bed           = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.bed.gz'

    blacklist          = "${projectDir}/assets/blacklists/GRCh37-blacklist.bed"

    spikein_fasta      = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.fa.gz'

    spikein_bowtie2    = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.tar.gz'



    minimum_alignment_q_score = 10

}

conf/igg_flow_tests/test_all_small.config

0 → 100644
+23 −0
Original line number Diff line number Diff line 
params {

    config_profile_name        = 'Test profile'

    config_profile_description = 'Minimal test dataset to check pipeline function'



    // Limit resources so that this can run CI

    max_cpus   = 2

    max_memory = 6.GB

    max_time   = 6.h



    // Input data

    input = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/samplesheet/test-GSE145187-all-small.csv'



    // Genome references

    fasta              = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20.fa.gz'

    bowtie2            = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-bowtie2.tar.gz'

    gtf                = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.gtf.gz'

    gene_bed           = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.bed.gz'

    blacklist          = "${projectDir}/assets/blacklists/GRCh37-blacklist.bed"

    spikein_fasta      = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.fa.gz'

    spikein_bowtie2    = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.tar.gz'



    minimum_alignment_q_score = 10

}

conf/igg_flow_tests/test_one_exp_all_igg_small.config

0 → 100644
+23 −0
Original line number Diff line number Diff line 
params {

    config_profile_name        = 'Test profile'

    config_profile_description = 'Minimal test dataset to check pipeline function'



    // Limit resources so that this can run CI

    max_cpus   = 2

    max_memory = 6.GB

    max_time   = 6.h



    // Input data

    input = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/samplesheet/test-GSE145187-1exp-alligg-small.csv'



    // Genome references

    fasta              = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20.fa.gz'

    bowtie2            = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-bowtie2.tar.gz'

    gtf                = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.gtf.gz'

    gene_bed           = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.bed.gz'

    blacklist          = "${projectDir}/assets/blacklists/GRCh37-blacklist.bed"

    spikein_fasta      = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.fa.gz'

    spikein_bowtie2    = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.tar.gz'



    minimum_alignment_q_score = 10

}

conf/igg_flow_tests/test_some_exp_all_igg_small.config

0 → 100644
+23 −0
Original line number Diff line number Diff line 
params {

    config_profile_name        = 'Test profile'

    config_profile_description = 'Minimal test dataset to check pipeline function'



    // Limit resources so that this can run CI

    max_cpus   = 2

    max_memory = 6.GB

    max_time   = 6.h



    // Input data

    input = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/samplesheet/test-GSE145187-varyingexp-alligg-small.csv'



    // Genome references

    fasta              = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20.fa.gz'

    bowtie2            = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-bowtie2.tar.gz'

    gtf                = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.gtf.gz'

    gene_bed           = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/hg38-chr20-genes.bed.gz'

    blacklist          = "${projectDir}/assets/blacklists/GRCh37-blacklist.bed"

    spikein_fasta      = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.fa.gz'

    spikein_bowtie2    = 'https://raw.githubusercontent.com/luslab/test-datasets/cutandrun/reference/genomes/e_coli_U00096_3.tar.gz'



    minimum_alignment_q_score = 10

}

CRA Git | Maintained and supported by SUSTech CRA and CCSE