Why Mate-Pair library?: long distance information to span contig gaps

Problems:

• low-complexity data
• lucigen mate-pair kit not very user friendly
• only 5-10% of your DNA gets circularized like it's supposed to
• only 10% of circular DNA contains junctions
• less than 1 ng of every microg is actual usable data

How to compensate:

• start with lots of DNA
• more efficient molecular biology
• use Tn5 (from Chris Vollmers)
  • recognizes and loads specific sequence (an adapter)
  • cuts the DNA and ligates an adapter in the same step
  • very efficient - all sheared DNA has adapters
  • add a biotinylated linker to the end of the adapters
  • 2×75 data

Other Issues:

• Lots of linker near the beginning of the reads
• those reads need to be filtered out
• We want AT LEAST 30bp of non-linker at the beginning

• For Tn5 data, linker sequence is more likely to be farther into the read
• That's a good thing! Almost always have at least 30bp before linker

What to do about read where you don't see any linker?

• might want to throw them out because we're not confident that they're actually mate pairs
• throws out tons of data if you are sequencing less than 2×300

How to avoid chimeric circular DNA?

• can't run it on a gel (circular dna smears)
• adjust insert size to ~4kb - chimeras are large and unlikely to circularize properly

2×75 data with long linker (60bp): we'll probably not read all the way through the linker, but we'll see bits of it.