How would you do?

Today I had my first quiz in Protein Structure and Function… my first quiz in Grad School. I think everyone failed it.
Mainly, I think everyone felt (as we had been warned) that this professor’s quizes have little, if anything, to do with what he actually taught.

*shrugs* So it goes, to quote Billy Pilgrim.

How would you have done? It was a 15 minute quiz:
You want to sequence a human protein you’ve identified as a spot on a 2d gel. You excise the gel spot and send it in for sequencing via Edman degradation. The Tech reports that there were problems, and you only get the first 4 residues of the protein.
1. How many possible sequences could the first 4 amino acid residues be? Make any assumptions necessary.
2. Given that there are ~40,000 human genes, what is your chances of correctly identifying the gene encoding your protein?
3. What were your assumptions in part (1), and how would they change your answers if they couldn’t be made?
4. Instead of Edman degradation, you try Mass Spec, but again the Tech has problems, and can only resolve a single, highly resolved and precise, molecular mass for your entire protein. What are your chances of correctly identifying your protein?

4 Responses to “How would you do?”

---

Almost sounds more like a statistics class or something than a “Protein Structure and Function” class…

What I’d do.. is figure since mayhaps everyone else failed, I’d try not to let it bother me so much but learn from it for the future.. 🙂 (Then again, I might not be that great a role model — I’m getting a bit burnt out on school/grades/etc as it is.) At least now you know what his quizes are like and a bit more of what to expect. Are these going to be weekly or something?

Teisha - September 30th, 2004 at 6:18 pm

---

Wow . . . it sounds painfully similar to my biochemistry class. I’m sorry it sucked. And, truly, it does sound a lot like a statistics course. All you can do is study what he teaches you and hope for the best.

Teisha, you’re a great role-model. 🙂 I always found that the less I focused on grades, the better my grades got. Learning is fun!

ShortSpeedFreak - September 30th, 2004 at 9:33 pm

---

Now to demonstrate why I am not a biochemist:

1. 20^4? 160,000
2. Er… pretty good? I mean, the above number is 4 times the number of genes, so if you’ve got that sequence you’ve essentially got it sussed.
3. I honestly have no frickin’ clue. Uh, that the gene wasn’t a highly conserved or recently duplicated homolog of another gene?
4. Being unfamiliar with the specificity of mass spec analysis, I’d say the chances are probably on the low end. You could maybe guess an overall content, but not sequence.

And now, how wrong was I? I need to get back in the lab, I swear.

Owen - September 30th, 2004 at 10:56 pm

---

Hey Owen-
SPOILERS
1. Absolutely correct- 20^4, or 160,000 possible sequences
2. You need to figure the odds of 2 genes encoding the same 4 amino acids, so odds of two identical starting sequences in 40,000 genes… I didn’t get thi part perfectly correct, but it’s incredibly small. So yeah- easy match….
3. Except that (a) not all 20 amino acid residues occur with equal frequency, so you can’t generalize that well (if your 4 sequence is mostl alanine, you may be fucked), and (b) you can’t assume that there is the same probabilitiy among the starting amino acids (for example, methionine is high and proline is practically unheard of), which is similar (maybe enveloped) by (a).
4. No one has figured out any better answer than “very high chance”. You should check out mass spec some time, Owen- because each atom has a very specific mass (like 8 or 9 digits of accuracy, or more), they can actually run the discovered mass against a simulation of the protein masses for known proteins, and it’s DAMN accurate a lot of the time. Unbelievable stuff.

Yeah- this was a probability test. I’m curious how the course will work out. *shrugs*

Paradoxdruid - September 30th, 2004 at 11:12 pm

---

Leave a Response (or trackback on your own site)

You must be logged in to post a comment.