Cross-hybridization

This is going to be a tough one, and I'm afraid I might end up with more questions than answers at the end of this blogpost. Either way I feel like I need to clear my mind on the matter of cross-hybridization.

After a long and nice summer holiday with my family, work didn't progress much, even though I had hoped that my external contacts would be able to do something after final submission of my microarray design. However, it turned out that everyone else was gone on their summer holidays too. I should have known.
Now everything is moving along nicely again, but the road up to the final design was certainly not a straight one, mainly due to cross-hybridization.

This issue has been a constant grain of sand in my eye since the day I set foot in a molecular lab, however, it got all the more profound when I started working with qPCR and gene expressions.
As of now, there's no satisfactory way of dealing with it, we just know that it's there and can do our best to account for it in our data analyses and interpretations.
My next project (after the microarray) will attempt to at least improve (decrease cross-hybridization) for some of the known and widely used oligonucleotides in marine cyanobacterial diazotroph research.

Cross-hybridization is basically hybridization of a primer or probe sequence with a sample sequence other than the intended target. Often it happens due to several sequences being highly similar, but also because the oligonucleotides used are not specific enough in their design.
What I mean with specificity in this case is that even though your oligonucleotide design is a perfect match to a certain region of a gene in a target organism, it could match almost, if not just as well, for a closely related species with the same gene. Moreover, even a close match (not perfect) could be enough for cross-hybridization to some degree. There are several reasons why mismatches in nucleotides still hybridize, like for example the nucleotide content of the target sequence (G-C-T-A) and the location of the mismatch in relation to the 5' and 3' ends.
To make matters even more complicated, the concept of species is kind of blurred when working with prokaryotes, where we often are interested in different strains of the same species. Even though they're systematically the same species, phylogenetically they're not, and potentially have different impacts on different ecosystems.

So running cross-hybridization tests where I use oligonucleotides designed for a particular target organism on a closely related strain or species to see if I get a signal. If I do get a signal, it is a cause for concern, but if I don't get a signal it is safe to say that the designed oligonucleotides are able to distinguish the target organism from at least that particular closely related strain or species which I tested against.
Of course it is impossible to test for every single possible cross-hybridization. There are simply too many unknowns involved and for some strains we don't even have any designed oligonucleotides yet. It is also possible that there are closely related strains or species out there in the ocean that we just don't even know exist.

An emerging number of marine cyanobacterial diazotrophs are just known by sequence, matched by phylogeny, in metatranscriptomic studies. My diatom diazotroph associations (DDAs) for example, are easily observable in a fluorescence microscope, so I can quickly look and see if they're there and which species and strain it is. That will initially be key to improving and optimizing oligonucleotide designs for the larger marine cyanobacterial diazotrophs. For the smaller we're somewhat dealing with a black box and I'm expecting a trial and error approach.
Up to this point we've mainly focused on the nifH gene for our target sequences, but as these organisms get fully sequenced, we might find alternative genes which are better suited for our purposes, based on relatively subtle genetic differences.