4.8 mergeOverlappingReads

Merging paired-end reads in BAM file

mergeOverlappingReads detects overlapping bases between paired‑end reads and soft‑clips all bases in the overlap. This ensures that no base is counted twice when both mates cover the same genomic position. By soft-clipping instead of removing the bases, they are ignored by downstream analyses while preserving the original read structure. It is advisable to run this task before any consecutive variant discovery or population genetic tool if reads weren’t merged yet.

4.8.1 Input

Required inputs :

--bam Input_bam_file.bam

Input BAM file.

Optional inputs :

--recal OR --RGInfo sample_RGInfo.json	Quality score recalibration file (see estimateErrors for further information).
--pmd OR --RGInfo sample_RGInfo.json	Post-mortem damage parameters (see estimateErrors for further information).

Specific Parameters :

--mergingMethod method	To specify the method to be used for merging. Options: middle, keepFirst, keepSecond, keepFwd, keepRev, random. Default = middle. See the merging methods table for additional information on all available merging methods.
--filterFragmentMismatches [float_value1,float_value2]	Remove all fragments whose ratio of number of mismatches to the combined mapped length outside of the specified range. Default = Don’t remove any reads regardless of number of mismatches.

• Filter parameters to apply specific filters for bases and reads.
• Engine parameters that are common to all tasks can be found.

4.8.2 Output

*_merged.bam	A BAM file with merged reads.
*_merged.bam.bai	Index files for merged BAM files.
*_filterSummary.txt	A text file with filter summary.

4.8.3 Usage Example

# Simulate a BAM File with paired end reads
atlas simulate --seqType "paired"

# Merge reads
atlas mergeOverlappingReads --bam ATLAS_simulations.bam --fasta ATLAS_simulations.fasta

4.8.4 Merging Methods

middle	Will keep half of the overlapping positions of each mate.
keepFirst	Will keep read of first mate at overlapping positions
keepSecond	Will keep read of second mate at overlapping positions.
keepFwd	Will keep read of forward strand at overlapping positions.
keepRev	Will keep read of reversed strand at overlapping positions.
random	Will keep random read for all overlapping positions.