Optimising PCR

I had a cool journey to Codfish Island recently, but as usual I didn't take any photos, so I'll wait a bit to see if I can use other people's before writing about that.

Right now, I am working towards building a family tree (AKA phylogeny) of Cantuaria. I have catalogued all of the specimens I collected; there are 189 spiders, and if their distribution is anything to go by, around 30 species. It's by no means the complete set, and I have a long way before I am a Pokémon master. But it's a start.

Each sample has to be dissected, have its DNA extracted, and then go through polymerase chain reactions (PCRs) until it is ready to be sequenced. Then I have a DNA sequence, which I fiddle with a bit and add to the collection. When I have a complete(ish) collection of sequences, I'll align them, look at different sketches of trees using them, and then argue with the program BEAST (Bayesian Evolutionary Analysis Sampling Trees) until it gives me a tree.

Genes
The genes I'm sequencing are EF1y, ITS and COI. EF1y and ITS are from the nucleus of the cell, and part of the spider's DNA that makes proteins to become bits of spider, or tell other bits of spider to do their jobs. COI is mitochondrial DNA. Mitochondria are little bacteria-like things in cells that help with respiration (converting glucose into energy using oxygen). They appear to have once been organisms in their own right, so they have their own genome with their own DNA. You inherit your mitochondria (along with their DNA) from your mother, as they float around in her eggs. Mitochondrial DNA evolves pretty quickly, so I'm using it to try and detect more recent changes in Cantuaria's lineage. The nuclear DNA evolves more slowly, so that hopefully it'll show me older changes. The changes I want to see are divergences - times when the lineage has split, for example into different species.

Optimising trials
At the moment I've dabbled a bit in COI, but my main efforts are directed at EF1y. Nuclear DNA is harder to amplify, but I think ITS will be harder to deal with because it has secondary structures (stuff I don't know much about but involves the DNA sticking to itself, I think).

First, I do a touchdown PCR. Then a nested PCR. Both are instructed in the paper that got this sequence from spiders before.

When I first had a go at amplifying EF1y, this happened:

The really bright smudge on the left (which is supposed to be a neat collection of lines) is a measuring tool called a ladder - it measures how long the piece(s) of DNA you amplified is. To the right of the ladder is the negative, which should not contain any DNA except the smudge at the bottom which is a primer dimer. To the right of the negative are the two spider samples. If the PCR had worked, I would have got 2 crisp white lines ("bands") roughly mid-way up the ladder. Instead I got 2 sets of horrible smudges in the wrong places. And an extra chunk of DNA that I didn't need (marked by the question mark). Despite the fact that I hadn't got anywhere near the result that I needed, I was thrilled with this, because it meant my extractions had DNA in them, and that I had managed to amplify something.

Since that PCR, I have changed a few things - DNA concentration, MgCl2 concentration, primer concentration, PCR annealing temperature, and added betaine. I also ran the gel for a bit longer to spread the bands (including the ladder) out. Yesterday I got this result:

 

You can hopefully see five sets of three columns - two tall and one short (the last set, on the bottom row, has no visible short column). The short columns are the negatives. The tall ones are Cantuaria DNA. Mostly, they have multiple bands that are a bit crisper than before.

The third set from the left used a reduced amount of MgCl2, and added betaine. They have two clearish bands in roughly the right place (I REALLY HOPE), and not much else - there is a faint band near the bottom in the left one, but nothing more than a smudge in the right. The last set of columns (using GoTaq Green instead of iTaq) also show only two clear bands, but they are much less crisp than the third set.

I'm pretty happy that I got only two (ish) bands, in the right place (I HOPE), and they're crisp (ish). They're not perfect but today I had a go at doing a cleanup run and then sequencing them. Tomorrow I should find out if I have amplified the right part of the right gene. Otherwise it's back to the drawing board.