The smear

My electrophoresis gel

This image represents progress. 

Recently, I've been trying to develop a quick and easy way to identify cryptic lineages of the coral Porites lobata. For the past few years, my team has researched Porites in Palau. We discovered that the corals living inside semi-enclosed lagoons have higher thermal tolerance than corals living outside the lagoons, and they're actually genetically distinct. Two of the genetic groups - we call them lineages - have higher thermal tolerance and are more common inside the lagoons. 

Every time we collect samples and want to sort the corals by lineage, it's an elaborate process. There are DNA extractions, then clean-up steps, then dilutions, then library prep, then sequencing, then bioinformatics - the whole thing takes months. Sure, the cost of sequencing has come down in recent years, but there's still a lot of time investment for genomic analysis. When you have a few hundred samples to weed through and just need a simple answer, it seems like the effort outweighs the result. 

So here I am. I have almost 200 corals in my freezer, and I need to identify each of them to lineage. I don't need anything else from them - just the lineage. In fact, sorting those coral samples into lineages is the first step in my analysis of reproductive isolation. The whole project will take months anyway, so I want to streamline and reduce the cost of that first step. 

My current strategy is to develop an RFLP: a restriction fragment length polymorphism. You may not know this, but there are a number of commercially-available enzymes that can cut DNA. Even better, each enzyme cuts DNA at a specific site - a specific sequence - so if you have two strands of DNA with different sequences, the enzyme will cut one but not the other. My task is to identify a particular spot in the DNA sequence for my four lineages that will work with one or more enzymes. If the enzymes cut one lineage but not another, or cut different lineages at different spots, I'll be able to tell those lineages apart. DNA that's been cut at different locations will be different lengths, so when I run the DNA out on an electrophoresis gel, each lineage will have a different pattern of bands. 

The hard part is picking the right spot in the DNA and picking the right enzymes. I started with a DNA locus called ITS2, mostly out of convenience, because I already had the correct primers in my lab and had honed the method for using them earlier this year. ITS2 is also supposed to be a highly variable section of DNA, so I thought it was likely to differ between lineages. I sequenced ITS2 for each lineage and compared the sequences - bingo. There were three enzymes that would cut the DNA in different patterns depending on the lineage. I ordered the enzymes and ran a test. 

It worked, kind of. I definitely got the enzymes to cut the DNA, but they might have worked a little too well. Check out my gel image above. In each column on the gel, you see a band at the top, a gap, and then a smear. The fact that there are two sections (the band and the smear) shows that the enzymes worked - the DNA was cut into at least two pieces. The part that worries me is the smear. The speed of DNA running through a gel depends on the length of the fragment, so a smear usually means you have a bunch of DNA fragments with similar but not identical lengths. It seems like maybe the enzymes went nuts and started chopping the DNA wherever they pleased. 

There are several things I can try next. I'll reduce the amount of time I incubate the DNA with the enzymes to hopefully restrict the cutting action. I'll run my gel at lower voltage for longer so the DNA fragments separate more along their lanes. I'll look back at my ITS2 sequences and see if there are other enzymes I missed that might work better. 

Science is always a slow, steady process. Right now, I'm in the trial-and-error phase, but I'm sure I'll have a breakthrough soon. Every project does.