Won't assembling reads to a reference remove SNPs and InDels?

[sdmoore]

When aligning reads to a reference to generate a config for SNP/InDel detection, won't the alignment necessarily remove the reads that have the desired information?

I see this as a commonly-listed step before searching for mutations.

Other threads I have read here (and papers referred to in them) refer to software that does the alignments (I have used Bowtie and Ugene), but I clearly see the absence in the resulting contigs of known antibiotic resistance markers that are in the genomes (large insertions). My filtered Illumina sets are paired (which don't seem to work in the alignments), but the individual libraries provide about 10 fold coverage with most reads >200.

[dpryan]

Most of the pipelines aren't intended for finding very large insertions of novel sequence. For those, just use a different method, such as this one.

[sdmoore]

Thanks for that paper, I had it in my folder, so I must have at least skimmed it.
I have not gotten decent assemblies using the paired end libraries (processed), I think they may all have to be the same length, which I didn't realize when I was filtering/trimming each.

The assemblers I have used don't spit out the orphans, which is exactly what I need. I will find a way to get those.

Thanks for the help.