Diversity

This article was originally published for Cambio Ltd. on 14-11-2012. Access it here.

In my original project description there is a very innocuous sentence that reads “rhizobia diversity will be assessed by PCR fingerprinting with REP/ERIC/BOX PCR”.

The problem is, it won’t.

Unfortunately, neither will RP01, RP04 or RP05 RAPD primers, or NodA or GyrB primers, as 2 years of tweaks and frustration have shown me. It would seem I’m at a dead end. So with a year to go, 1000 strains and a £2,000 budget I am faced with the dilemma of how to address my project title of “Rhizobia Diversity in Farm Soils”. Hum.

One of the main problems of this is the more I delve into the topic, the more I wonder what ‘diversity’ means. A quick look in my Chambers dictionary yields the definition:

n diverˈsity state of being diverse; difference; dissimilarity; variety.

But how much difference? How much dissimilarity? How much variety?

I’m beginning to see why I will be a doctor of philosophy…

My original hope was to use banding patterns to identify my strains. My NodA assay very reliably gives me two very different banding patterns for the two bacteria I work with: Rhizobium leguminosarum bv. trifolii and Sinorhizobium meliloti. But it doesn’t give me any finer resolution, and I can already identify with some reliability to the species level because of the plant I’ve isolated them from in the first place. And I must say looking at a plant is a much quicker way of discerning that level of diversity!

So discernment to a species level is pretty easy when you’re coming at it from the direction I am. If it comes off an alfalfa plant, chances are it’s S. meliloti. No need to spend a week showing that molecularly. But my project is aimed at discerning more subtle differences that might have been responsible for crop failures in a previous project: perhaps changes in a gene that affects nodulation ability under low pH, or desiccation, or Ca deficiency. Or something that improves N-fixation ability. Or, or, or… something!

So how do I find that difference?

There are currently 92 species of ‘true’ rhizobia in 12 genera, though it sometimes seems that these classifications change on a bi-weekly basis. I think I’m right in saying that they’re based on 16S rRNA profiles: differences in sections of housekeeping genes that are similar, but whose differences merit their classification as a separate species.

This week I had my first lesson in metagenomics and phylogeny, a combination of brain-crunching pain and excitement at delving into the private lives of my bacteria. Using the 16S rRNA sequence data for 4 wild-type strains of S. meliloti, including its sister strain S. medicae, I crunched them in BioEdit and found… very little difference.

16S is clearly not the gene I will be using for this method.

There are other options though, and with the age of next-generation sequencing upon us it is much easier to analyse diversity at a very fine level using this method: it is clearly the way forward. The problem is the cost. I have almost 1000 strains, far more than I can afford to sequence. My alternative is RFLP, which is cheaper but a nightmare to analyse for that many samples, and won’t give me such fine detail. So I am left with a dilemma that must be familiar to a lot of university-based researchers: how do I squeeze the best value out of my funding? How can I get some original, publishable, rigorous results? At this point I’m considering paying for the sequencing myself!