Optimized compare_ref function from loop logic to dataframe logic

[TrangNg-Th]

Description

This pull request optimizes the compare_ref function for faster run of bam merging step after creating mutated bam file. In this version, I

• Modified the function to use dataframe logic instead of loop logic, using the pandas package to handle and compare references faster.
• Added a bamsurgeon.def Singularity definition file, allowing users to build a Singularity image for executing the code.
• The command to build the Singularity image is: singularity build bamsurgeon.sif bamsurgeon.def

Additional Dependencies

• pandas
• (optional) Singularity

Benefits

• Significantly reduced runtime by processing references more efficiently.
• Improved code readability and maintainability by using dataframe operations.

Benchmark

Test Setup:

• Donor BAM file size: 3GB
• Reference FASTA file: Mmul 8.0.1, consisting of 284,727 contigs (including 22 chromosomes)
• Single SNP

Performance Improvement:

• The addsnv.py script execution time has been reduced from approximately 4 hours to 2-3 minutes with the optimized compare_ref function.