Alignment to a set of custom reference sequences along with standard genome reference

[eeyun]

Hi there - any help appreciated here, I'm a relative newbie to NGS evaluation.

I'd like to be able to detect low level deletions of 15bp (or more) in an amplicon-enrichment based assay. We are using Novoalign for alignment, followed by samtools and varscan for variant ID. This all works find up to ~6bp.

I have tried pindel etc. with limited success. As the deletions we are looking for are relatively recurrent, I was thinking of trying a synthetic reference approach. I'm unclear how I can integrate this alignment strategy with our alternative standard reference alignment.

I'd prefer to create a single fasta file as a reference, but for all of the short synthetic deletion sequences I append (one per record), I'd like to preserve their original chromosomal position etc.

Am I dreaming?

Any help or other recommendations much appreciated!

[swbarnes2]

You can of course make a reference file that has multiple sequences in it, so you can append previously discovered deletions into your fasta, and then if your reads align to those sequences, you will see them.

But making their coordinates match is not feasible, and I don't see why it's necessary.

[eeyun]

You can of course make a reference file that has multiple sequences in it, so you can append previously discovered deletions into your fasta, and then if your reads align to those sequences, you will see them.

But making their coordinates match is not feasible, and I don't see why it's necessary.

Thanks for your input. I am hoping to preserve the coordinates to enable visualization and annotation without too much deviation from our standard practice.

[KaiYe]

Hi there - any help appreciated here, I'm a relative newbie to NGS evaluation.

I'd like to be able to detect low level deletions of 15bp (or more) in an amplicon-enrichment based assay. We are using Novoalign for alignment, followed by samtools and varscan for variant ID. This all works find up to ~6bp.

I have tried pindel etc. with limited success. As the deletions we are looking for are relatively recurrent, I was thinking of trying a synthetic reference approach. I'm unclear how I can integrate this alignment strategy with our alternative standard reference alignment.

I'd prefer to create a single fasta file as a reference, but for all of the short synthetic deletion sequences I append (one per record), I'd like to preserve their original chromosomal position etc.

Am I dreaming?

Any help or other recommendations much appreciated!

Pindel should find large deletions. can you pass your data to me to take a look?

Kai

[shi]

Hi there - any help appreciated here, I'm a relative newbie to NGS evaluation.

I'd like to be able to detect low level deletions of 15bp (or more) in an amplicon-enrichment based assay. We are using Novoalign for alignment, followed by samtools and varscan for variant ID. This all works find up to ~6bp.

I have tried pindel etc. with limited success. As the deletions we are looking for are relatively recurrent, I was thinking of trying a synthetic reference approach. I'm unclear how I can integrate this alignment strategy with our alternative standard reference alignment.

I'd prefer to create a single fasta file as a reference, but for all of the short synthetic deletion sequences I append (one per record), I'd like to preserve their original chromosomal position etc.

Am I dreaming?

Any help or other recommendations much appreciated!

Hi @eeyun,

You may try the Subread aligner (http://subread.sourceforge.net) which can detect up to 16 bp indels.

Cheers
Wei