AGBT talk: Daniel Neafsey, Broad Institute of MIT and Harvard

Title: Hybrid selection for sequencing pathogen genomes from clinical samples

Sample bottleneck for pathogen sequencing applications.  Host DNA monopolizes any sample, but culturing pathogens takes 6 weeks and can introduce bias.  Instead, use Hybrid selection.

Example: Plasmodium falciparium Malaria, but generalizes to any pathogen. (Some background presented.)

  • Seq first in 2003,
  • AT Rich: 81%!
  • hard to sequence

“Agilent’s Sure select system” is new name of Hybrid Selection.

  • Use bait, wash away non target DNA, then wash way rest, unhybridize.

40-fold enrichment when done on parasite DNA.  But, starting from blood spots, you don’t always have a lot.  However, no penalty for Whole Genome Amplification. (Same results as for straight blood.)

Technology helps remove human contamination, and gives much better data on regions covered by baits.

New approach: Whole Genome Baits (Rogov & Melnikov).

Whole Genome -> adaptors, (T7 tail) -> bait.

Whole genome bait results are comparable to synthetic baits, and again, no penalty for simulated clinical samples.

Coverage Similar to pure parasite DNA.  There are peaks and troughs, but comparable to pure DNA.  Dips correspond to AT rich regions, which are hard to sequence anyhow, and can’t be improved much over what you get with regular sequencing.

Tested on real clinical sample – blood spot, one year old.  Amplified, then whole genome selection.  Used HiSeq, and got 5.9% plasmodium DNA.

Figure of “Accurate SNP calling from Hyb-Sel data.”  Not just picking up parts that look like ref plasmodium, but also finding other stuff, which influences SNP calling – but still reasonable.

Other examples, with 50x enrichment.  Did other examples from 12 clinical samples, again saw good enrichment.

Summary: This is a good method for enriching pathogens, which helps reduce the amount of sequencing needed – and that means cost is also reduced.  Good for everyone.

Used on malaria for a model, but could work well on other pathogens.  Will enable sequencing of clinical samples collected in drug & vaccine trials.

2 thoughts on “AGBT talk: Daniel Neafsey, Broad Institute of MIT and Harvard

  1. Pingback: Tweets that mention AGBT talk: Daniel Neafsey, Broad Institute of MIT and Harvard | --

  2. Pingback: The roundup of twittered updates for 2011-02-05 « The Hyphal Tip

Leave a Reply

Your email address will not be published. Required fields are marked *