Commit b67c7e4d authored by peguerin's avatar peguerin
Browse files

format snakemake

parent c28064aa

Snakefile

+1 −2
Original line number Diff line number Diff line
@@ -7,7 +7,6 @@ configfile: "config.yaml" 
rule all:

    input:

        expand("{folder}/{fastqf}_R1.fastq.gz", fastqf=config["fastqFiles"],folder=config["fastqFolderPath"]),        

        expand("{folder}/{barcode}.dat", barcode=config["barcodeFiles"],folder=config["fastqFolderPath"]),

        expand('assembled/{run}/{run}.fastq', run=RUNS),

        expand('assembled/{run}/{run}.ali.fastq', run=RUNS),

        expand('assembled/{run}/{run}.ali.assigned.fastq', run=RUNS),

config.yaml

+11 −11
Original line number Diff line number Diff line
@@ -15,17 +15,17 @@ fastqFiles: 
  - 161124_SND393_A_L005_GWM-858

  - 161124_SND393_A_L005_GWM-859

barcodeFiles:

  - MB1016K_Teleo

  - MB1016L_Teleo

  - MB1016M_Teleo

  - MB1016N_Teleo

  - MB1016O_Teleo

  - MB1016P_Teleo

  - MB1016Q_Teleo

  - MB1016R_Teleo

  - MB1016S_Teleo

  - MB1016T_Teleo

  - MB1016U_Teleo

  - 161124_SND393_A_L005_GWM-849 : MB1016K_Teleo

  - 161124_SND393_A_L005_GWM-850 : MB1016L_Teleo

  - 161124_SND393_A_L005_GWM-851 : MB1016M_Teleo

  - 161124_SND393_A_L005_GWM-852 : MB1016N_Teleo

  - 161124_SND393_A_L005_GWM-853 : MB1016O_Teleo

  - 161124_SND393_A_L005_GWM-854 : MB1016P_Teleo

  - 161124_SND393_A_L005_GWM-855 : MB1016Q_Teleo

  - 161124_SND393_A_L005_GWM-856 : MB1016R_Teleo

  - 161124_SND393_A_L005_GWM-857 : MB1016S_Teleo

  - 161124_SND393_A_L005_GWM-858 : MB1016T_Teleo

  - 161124_SND393_A_L005_GWM-859 : MB1016U_Teleo

illuminapairedend:

  s_min : 40

good_length_samples:

rules/step1.sf

0 → 100644
+68 −0
Original line number Diff line number Diff line 
configfile: "config.yaml"

RUNS, = glob_wildcards('raw/{run}_R1.fastq.gz')

BARCODES, = glob_wildcards('barcodes/{barcode}.dat')

DICBARCODES={}

i=0

for bc in BARCODES:

   DICBARCODES[RUNS[i]]="barcodes/"+bc+".dat"

   i=i+1

#print(DICBARCODES)

rule all:

    input:

        expand('assembled/{run}/{run}.fastq', run=RUNS),

        expand('assembled/{run}/{run}.ali.fastq', run=RUNS),

        expand('assembled/{run}/{run}.ali.assigned.fastq', run=RUNS),

        expand('assembled/{run}/{run}.unidentified.fastq', run=RUNS),

        expand('log/remove_unaligned/{run}.log',run=RUNS),

        expand('log/illuminapairedend/{run}.log',run=RUNS),

        expand('log/assign_sequences/{run}.log',run=RUNS),

        expand('log/split_sequences/{run}.log',run=RUNS)



### Paired end alignment then keep reads with quality > 40

rule illuminapairedend:

    input:

        R1='raw/{run}_R1.fastq.gz',

        R2='raw/{run}_R2.fastq.gz'

    output:

        fq='assembled/{run}/{run}.fastq'

    log:

        'log/illuminapairedend/{run}.log'

    params:

        s_min=config["illuminapairedend"]["s_min"]

    shell:

        '''illuminapairedend -r {input.R2} {input.R1} --score-min={params.s_min} > {output.fq} 2> {log}'''



### Remove unaligned sequence records

rule remove_unaligned:

    input:

        fq='assembled/{run}/{run}.fastq'

    output:

        ali='assembled/{run}/{run}.ali.fastq'

    log:

        'log/remove_unaligned/{run}.log'

    shell:

        '''obigrep -p 'mode!=\"joined\"' {input.fq} > {output.ali} 2> {log}'''



### Assign each sequence record to the corresponding sample/marker combination

rule assign_sequences:

    input:

        'assembled/{run}/{run}.ali.fastq',

        lambda wildcards: DICBARCODES[wildcards.run]

    output:

        assign='assembled/{run}/{run}.ali.assigned.fastq',

        unid='assembled/{run}/{run}.unidentified.fastq'

    log:

        'log/assign_sequences/{run}.log'

    shell:

        '''ngsfilter -t {input[1]} -u {output.unid} {input[0]} --fasta-output > {output.assign} 2> {log}'''



### Split the input sequence file in a set of subfiles according to the values of attribute `sample`

rule split_sequences:

    input:

        'assembled/{run}/{run}.ali.assigned.fastq'

    params:

        'samples/{run}_sample_'

    log:

        'log/split_sequences/{run}.log'

    shell:

        '''obisplit -p "{params}" -t sample --fasta {input} 2> {log}'''

rules/step1.smk

+1 −1
Original line number Diff line number Diff line
@@ -31,7 +31,7 @@ rule remove_unaligned: 
rule assign_sequences:

    input:

        'assembled/{run}/{run}.ali.fastq',

        lambda wildcards: DICBARCODES[wildcards.run]

        lambda wildcards: config["barcodeFiles"][wildcards.run]

    output:

        assign='assembled/{run}/{run}.ali.assigned.fastq',

        unid='assembled/{run}/{run}.unidentified.fastq'