Friday, April 29, 2005

Science is all crap.

Well, it has been a few days and I have no new data. I forgot what this was like: as a postdoc, stuff has generally worked for me. I've had some dry spells, but I usually always have something cooking. A few days of solid failure brings me back to grad school... ahh, grad school. What a miserable time that was!

What I can report is that I can't clone this mystery protein, now that I know what it supposedly is. I can detect the transcript, no problem, using a commercial quantitative PCR kit (of course the sequences of their probes are proprietary). But when I go to amplify it to clone it, nada. Nothing. So I don't know if the Q-PCR is crap (probably) or if it's all my fault (quite possibly). Probably both!

Meanwhile, I have been trying to do sequence alignments. Now mind you, I'm familiar with proteins that are not well-conserved, and I have done quite a few alignments using CLUSTALW, and by hand. So my mystery protein is supposedly related to a well-conserved family of proteins, and it comes up in BLAST with this similarity. But when I do alignments with various family members, it is not at all obvious that this protein actually belongs to this family. Depending on which member I choose, I get a totally different alignment. I mean, TOTALLY. If there are really 20 conserved residues, as the literature says, in this family, shouldn't those at least come up every time? Noooooo....

Anyway I guess I am off to prep more RNA, because I'm worried my cells might not last until tomorrow if I blow it off, and possibly stain some coverslips, mostly because I'm worried I will be too lazy to come in and do it tomorrow.

Hard to want to come in for more punishment on the weekend.


At 11:21 AM, Anonymous Anonymous said...


when you do pairwise alignments of your protein and one of the members in the family, there is no guarantee that all the 20 conserved residues will show up in the alignments. To get those conserved residues, you need to do a multiple sequence alignment (taking a few more sequences, hoping to nail down the really conserved sequences), or run a PSI-BLAST to get the profile. If you still get nothing, well something weird is going on.

At 12:44 PM, Blogger Adam Solomon said...

If there's no documented explanation for your problem, you can find out what's causing it. And that's a good thing.


Post a Comment

Links to this post:

Create a Link

<< Home