Step right up; predict a promoter and win...

Win what? Well, nothing more than a few months of hard work, hopefully followed by a nice wee paper...

We're in the mood for science on akatsuki talks rot. Perhaps this has come about since I've acknowledged that I will always be a mediocre blogger, but I could become a better science communicator if I practice writing about things I work on more. And all that stuff about the bloody Viagra-jet lag story (and the media coverage of it) made me realise there is a middle-ground for scientists in the know to discuss hot topics without resorting to the extremes of technobabble (which we're used to applying in our daily work) nor the overly condensed data-free reportage by the media.¹

I've toyed with the thought of using this blog to practice writing mini-reviews of interesting published scientific studies, but have been to lazy to date. This may continue, or it may not. But given that nobody actually reads this crap, it doesn't matter. (Hi P! You still reading this surreptitiously? I have a site counter, you know...)

At any rate, since my laptops have a tendency to die after a few years, and I'm not very good at backing up non-essential data, here follows an email I spent a good 15-20 minutes on this evening to a colleague. I think I was being Ms Good Postdoc, but I often fear that I don't give good advice. Time will tell.

The premise: a friend of a colleague needed a starting point for a promoter assay s/he is planning.

My reply to the request, slightly edited:

Hi [colleague],
There are several approaches to finding the promoter of a gene. The first would be to do a quick literature search to see if anyone has already determined the components of that gene's promoter.
If not, the second approach would be to retrieve the gene information from a genomic database (like Ensembl or the UCSC genome browser). These two sites are particularly useful because they allow very quick and easy comparisons across species (e.g. human/mouse/rat/fish). Such comparisons often show what sequence is most conserved: usually the exons and any regulatory elements that are essential (like promoter elements). This only works some of the time and will only highlight very conserved regions. There will be species-specific differences in promoter sequence.
The third, slightly more laborious approach, is to take several kilobases of sequence upstream of the first known exon and perform a promoter prediction. This has become relatively easy on the internet. A quick Google search found these possibilities.
Of course, the difficulty is in choosing one. I cannot recommend one over the other because I haven't tested them myself. There is one particular site I know of that is quite easy to use: Promoter 2 prediction. It looks for known promoter motifs based on already studied PolII promoters.
After that, your friend will have to make the decision of how much of the promoter to clone, and how many different construct lengths he/she wants to try. I guess it will also depend on the restriction sites available in the sequence and how big the sequence is.
As for getting the DNA, it can either be cloned by PCR if it's very small (around 10kb), or it can be "cut" out of a BAC (bacterial artificial chromosome) or PAC or cosmid (large-capacity plasmid) that already contains the genomic region (can be bought from places like the BAC/PAC resource or Invitrogen or Roswell depending on species). The restriction digest method is preferred because it will have no additional errors, and it is also easier to change the length of the promoter region by choosing different restriction enzymes. Some people even use the entire BAC and make small deletions to eliminate predicted regulatory elements.
I hope this is will be helpful in getting your friend started. The real difficulty will be making the construct and testing it; I wish your friend the very best of luck with this!

After sending it, I read what I wrote² and thought: goddamit, it's all too simplistic. They probably already know all this crap and were probably looking for something concrete. Like a real protocol. Unfortunately, for this and quite a lot of the other molecular biology I do, I do it on the hoof. At some point in the last 10 years, I followed somebody else's protocol (usually the postgrad/postdoc/technician I was asking advice from at the time). Then found alternatives, some of which I now use routinely, some of which I was too lazy to follow-up. Then I switched specialities. Twice. Now, I am used as a mol bio repository by my (ex-)lab. They sometimes even think of me as an expert. But a real expert wouldn't feel so fraudulent in giving advice, would she? Self-doubt creeps in... Oh what if I've screwed up somebody's project giving bad advice?

¹ I'm in no way implying real media gets it wrong (although some, like the gawd-awful Daily Mail, get it wrong and deliberately so!); just that they make the assumption that folk just want the headlines. Which may be true. But not this particular folk or her ilk.

² And after writing this post, I read it and thought: WTH? Why can't you stay on topic? I give up. It's time for dinner.