I've gotten a few questions recently about doing phage genome sequencing on Illumina sequencers, and we ourselves have been pumping genomes through our Illumina MiSeq and loving it. But I wanted to issue a word of caution to anyone considering Illumina for their phage genome needs: don't use the Nextera (or other "tagmentation") kits for library prep.
But isn't tagmentation awesome?
Yes.
A traditional next-generation sequencing library prep workflow looks something like:
- Fragment DNA by sonication or enzymes and cleanup.
- End-repair the fragmented DNA and cleanup.
- Size-select appropriate fragments.
- Ligate adapters and barcodes and cleanup.
- Amplify by PCR and cleanup.
A tagmentation workflow uses a transposon to simultaneously fragment the DNA and attach adapters, so it's more like:
- Tagment DNA and cleanup.
- Amplify by PCR and cleanup.
That sounds (and is!) a lot easier. Plus, you can use much smaller amounts of DNA for a tagmentation library prep. So, yes, tagmentation is awesome!
So why shouldn't I use it?
Simply: phage genome ends. Many phage genomes have defined physical ends, and if you're doing serious phage sequencing you probably want to know where those ends are. If you use a tagmentation approach, you'll have two problems related to ends.
- You won't see large buildups of reads that indicate the defined physical ends.
- Coverage will peter out and possibly drop to zero right at the physical ends.
This is because you can't insert a transposon between a nucleotide and...nothing. So you'll almost never get those last few bases near the genome ends, and even if you do, how will you know you've reached an end?
The "fix" for this is to do PCR and Sanger to get your genome to "wrap around" into a circle again. That's not too hard, but it's harder than nothing, which is what you generally have to do if you use a different library prep approach.
So far this year we've sequenced over 100 phages with defined ends, and having to do PCR and Sanger for each of those would've been a headache. Instead, we've been using the TruSeq Nano DNA prep kits for our libraries, and all of those ends have been obvious in the sequencing output. (And by the by, we don't have a fancy sonicator...we've been using the enzyme Shearase from Zymo, and it's working very well.)
To sum up
Tagmentation workflows, like Nextera, make for easy library prep but essentially guarantee you'll have to do post-sequencing bench work for any phage with defined physical ends.
Shearing workflows, like TruSeq, take slightly longer to do library prep but nearly 100% of genomes will require no post-sequencing bench work.
We've been doing 32 phage genomes per MiSeq run, and prepping 32 TruSeq libraries takes one long day, so while it's not quite as easy as tagmenting, it's also not remotely prohibitive. So my strong recommendation for phage genomes is to avoid tagmentation.