When Hard clipping is used instead of the Soft clipping ?

Question

[2012secondseason]

Hi all. The documentation for Samtools is minimal at best. I'm still confused on the concept of a clipped read.

I know that there are many explanations here that soft clipped bases are present in SEQ unlike hard clipped bases. But it's insufficient. In what cases, use H instead of S ??

Thanks.

Answer 1

[Pierre Lindenbaum]

in bwa, HARD clipping is used for supplementary reads.

Information (full read sequence) is still available in the primary alignment.

Answer 2

[Devon Ryan]

This is purely up to the program creating the file, which is why this isn't described in the samtools documentation. A good example of a time when most programs would hard-clip a read is if they remove known adapter sequence. It's then not worthwhile including that sequence in the output. One could also think of hard-clipping regions when they're part of supplemental alignments (I've seen both hard and soft-clipping for that, though the former is probably more useful).

[Carambakaracho]

2012secondseason, for example, BWA mem gives you some control how you'd like to have the clipped sequences represented - you can toggle between soft and hard clipping using the -Y option. There used to be an option to remove soft clipping altogether, but that seems to have changed its meaning (at least comparing the online manual to the command line description)