The importance of LB field in @RG tag for BreakDancer analysis

Question

[Nicola Casiraghi]

I have WGS paired-end data and I want to use BreakDancer to call structural variants.

Both my case and controls bam files do not have, in their header, the LB field in @RG tag and BreakDancer reports this lack. I can re-header bam files and add the missing LB easily, however I cannot understand how this information is important for BreakDancer and which are the recommended criteria to add them, i.e. a unique LB for all samples or specific for each of them? Is it referred to the Library Preparation Kit?

Could you provide me some hints to solve the problem? Thank you

Answer 1

[Devon Ryan]

Just use picard's AddOrReplaceReadGroups tool to replace the read groups. Assuming that each sample was sequenced only once, just set the library (LB) field to be the same as the sample (SM). So java -jar AddOrReplaceReadGroups.jar I=original.bam O=new.bam SM=sample1 LB=sample1 ....

Answer 2

[Istvan Albert]

The LB field is used to identify molecular duplication issues and other sequencing imbalances. For example if the same library is sequenced on different lanes some tools can use this to identify sequencing related artifacts.

[Nicola Casiraghi]

Thank you Istvan.

Answer 3

[ernfrid]

I believe BreakDancer uses the LB tag in two major ways:

1. To maintain information about that library's insert size distribution. Differences between independent library preparations can intentionally or unintentionally occur and you'll get optimal SV calling by indicating to the software that there are different libraries present.
2. To increase confidence in a call. SV supporting reads from different libraries increase the score.

[Nicola Casiraghi]

Thank you. However I am still quite confused about the problem. The LB tag is somehow referred to the Library Preparation kit? Could you show me an example of a valid LB tag exploited by BreakDancer? Many thanks again.