Bwa - Map Reads To Reference¶
For input preprocessed reads bwa finds the most similar genomic region in the provided reference genome, using bwa mem algorithm.
Location
- Filepath: <SnakeLines_dir>/rules/paired_end/mapping/mapper/bwa.snake
- Rule name: bwa__map_reads_to_reference
Input(s):
- r1: gzipped fastq file with left reads, e.g. ‘reads/%s/{sample}_R1.fastq.gz’
- r2: gzipped fastq file with right reads, e.g. ‘reads/%s/{sample}_R2.fastq.gz’
- index: reference index (created by rule bwa__prepare_index), e.g. ‘reference/{reference}/bwa_index/{reference}.bwt’
Output(s):
- bam: mapped read in BAM file, e.g. ‘mapping/{reference}/original/{sample}.bam’
Param(s):
- index: name of reference, technically filename’s path prefix, e.g. ‘reference/{reference}/bwa_index/{reference}’
- additional: additional params
- concordant: more strict conditions for scoring options should help to map only concordant reads
Bismark - Map Methyl Seq Reads To Reference¶
For input preprocessed reads treated by Bisulfide Bismark finds the most similar genomic region in the provided reference genome.
Location
- Filepath: <SnakeLines_dir>/rules/paired_end/mapping/mapper/bismark.snake
- Rule name: bismark__map_methyl_seq_reads_to_reference
Input(s):
- r1: gzipped fastq file with left reads, e.g. ‘reads/%s/{sample}_R1.fastq.gz’
- r2: gzipped fastq file with right reads, e.g. ‘reads/%s/{sample}_R2.fastq.gz’
- ct_index: CT reference index (created by rule bismark__prepare_index)
- ga_index: GA reference index (created by rule bismark__prepare_index)
- ref: reference genome, i.e. ‘reference/{reference}/{reference}.fa’
Output(s):
- bam: mapped reads in BAM format
- alignment_report: Control report with basic mapping statistics generated by Bismark
Param(s):
- aux_bam: bam file that have preset name by Bismark. Has to be renamed to match downstream analysis.
- aux_alignment_report: alignment report that have preset name by Bismark. Better to move to separate folder to keep bam directory clean
- bam_dir: directory with the output bam file
- reference_dir: directory with reference fasta, i.e. ‘reference/{reference}’
Bowtie2 - Map Reads To Reference¶
For input preprocessed reads bowtie2 finds the most similar genomic region in the provided reference genome.
Location
- Filepath: <SnakeLines_dir>/rules/paired_end/mapping/mapper/bowtie2.snake
- Rule name: bowtie2__map_reads_to_reference
Input(s):
- r1: gzipped fastq file with left reads, e.g. ‘reads/%s/{sample}_R1.fastq.gz’
- r2: gzipped fastq file with right reads, e.g. ‘reads/%s/{sample}_R2.fastq.gz’
- index: reference index (created by rule bowtie2__prepare_index), e.g. ‘reference/{reference}/bowtie2_index/{reference}.1.bt2’
- ref: reference genome, e.g. ‘reference/{reference}/{reference}.fa’
Output(s):
- bam: mapped read in BAM file, e.g. ‘mapping/{reference}/original/{sample}.bam’
Param(s):
- index: name of reference, technically filename’s path prefix, e.g. ‘reference/{reference}/bowtie2_index/{reference}’
- additional: additional params
- concordant: only concordant reads