Best practice for preprocessing the bam file for copy number estimation

[ty23991]

Hi all,

There are known limitations involved in the copy number estimation from exome sequencing (Tumor/Normal).

For instance the following paper provides some insights.
Teo et al. Statistical challenges associated with detecting copy number variations with NGS. Bioinfo. 2012


I have two questions:

1. In case of T/N paired analysis, what is the best practice for data preprocessing ?
Currently available duplicates removal approaches have limitations in discriminating true PCR duplicates and genuine biological sequences.

What is the best approach for duplicates removal?
What is the best suitable aligner for preparing the bam file for CNV estimation ?

2. In case normal reference bam is unavailable, what are the best recommended tools for copy number analysis in case of exome sequencing ?


Thanks

[dmb107]

Hi all,

What is the best approach for duplicates removal?
What is the best suitable aligner for preparing the bam file for CNV estimation ?

Thanks

I am also interested in best practices for preprocessing bam files in preparation for CNV estimation.

In addition to these questions, should reads with low MAPQ be filtered out?

Thanks, all.

[maxsalm]

Hi there,

personally I have had success with BWA/BWA-MEM followed by Picard's MarkDuplicates. N.B. It is important to align to the FULL genome (not just canonical chromosomes).