Panini Linguistics Olympiad: Genes from Space

Question

Problem:

Alien Protein Codes

Dr. Muzabique figured out how to manipulate six alien proteins to do three functions: construct, cut, and pack. The alien genetic code used to manipulate the proteins was found to have 6 "alphabet letters": A, T, G, C, D, N. Only he knew the top-secret algorithm, and two weeks ago, he was mysteriously found dead in his apartment.

The top-secret project is now in a fund crunch but the government decides to give it one more chance if the algorithm can be decoded. They send you the only page that was recovered from Dr. Muzabique's notebook, with one set of input instructions and output genetic codes, and ask you to decode the algorithm:

"Today was a great breakthrough! I found that large and small proteins inherently behave in different ways. The way they operate under different cellular functions..."

Input	Output
Celebi, construct Articuno	CNDATACACCNDTGTATTCGGGATGCNDTGCCCACCCCNNNCNDCTTTTCCACTNANCND
Articuno, construct and cut Terrakion	CNDTGTATTCGGGATGCNDACTGCAGGGACTCCNDGCATGACATCNNNNNNCNDTGCCCACCC
Terrakion, construct and pack Shaymin	CNDACTGCAGGGACTCCNDCAGGCTCNDGCATGACATCNNNNANCNDCTTTTCCACTNNNCND
Cressela, construct and cut and pack Azelf	CNDCCATATGATGCGACNDGTTGATCNDGCATGACATCNNNNANCNDTGCCCACCCCNANCND CTTTTCCACTNNNCND
Celebi, cut Cressela	CNDATACACCNDCCATATGATGCGACNDGCATGACATCNANNNNCNDCTTTTCCACTNANCND
Shaymin, pack and construct Azelf	CNDCAGGCTCNDGTTGATCNDTGCCCACCCCNANCND
Celebi, construct and cut and pack Cressela	CNDATACACCNDCCATATGATGCGACND
Terrakion, pack Articuno	CNDACTGCAGGGACTCCNDTGTATTCGGGATGCNDCTTTTCCACTNNNCND

Assignment:

Preserve the legacy of Dr. Muzabique! Explain how the algorithm converts input commands into output genetic codes.

The Request

Could someone please check to see if my solution below is correct? Thank you!

My Solution

Sections of the gene sequence are separated by the codon CND, like below:

The underlined parts are the proteins involved. The first underlined part is the protein doing the action(the first protein mentioned) while the second underlined part is the protein being acted upon(the second protein mentioned). I've color-coded the underlined parts to differentiate between proteins.

The other sections are the actions done. As hinted in Dr. Muzabique's notebook, the length of the underlined protein sequences affects the action sequences. If the first protein(the acting protein) is long, then the action sequences will represent the actions in the command. If the acting protein is short, then the action sequences will represent the actions not in the command.

The action sequences also seem to end in either NNN or NAN, or a sequence of the two. I'll call these sequences "penguins"(because why not) and the other part of an action sequence the polar bears. This part seems to depend on the size of the proteins. The third gene output gives us the key between the actions and the polar bears. For the action sequence for constructing(the polar bear is GCATGACATC), it seems that it has N_NN_N as the penguin. It also seems like the blanks represent the first and second protein lengths, in order, where it is A if the protein is short and N if the protein is long. For the action sequence for cutting(polar bear: TGCCCACCCC) the penguin is N_N, and depends on the second protein: if it is short then the blank is A, if it is long then the blank is N. For packing(polar bear: CTTTTCCACT) it is like the cutting penguin except it depends on the first protein.

^{Problem images: (1), (2)}

Source link: https://ltrc.iiit.ac.in/nlpmt/plo/#/resource.

Answer 1

Only whoever created the problem can say for sure, but your solution seems to match everything that's in the problem quite satisfactorily and I would expect it to get all or almost all the marks available. But, for the avoidance of doubt, I have never either entered or marked a linguistics olympiad, and I am a mathematician rather than a linguist, and there may well be (e.g.) particular conventions that I'm unaware of or subtleties that I've missed but someone more expert wouldn't.

Some things I would consider writing a bit differently (though I do not claim that this would make any difference to scores in the actual contest -- but there's one thing I do consider an actual omission, albeit a minor one, which I've put in bold below):

• Rather than saying in two separate places that NAN indicates a small protein and NNN a large one, I would say it just once. Something like: "Each action sequence ends with an indication of the sizes of one or both proteins involved. In each case NAN indicates a small protein and NNN a large one. The construct action has such size-markers for both proteins, subject first and then object. The pack action has a size-marker only for the subject. The cut action has a size-marker only for the object."
• You have left it implicit what "large" and "small" mean. Of course there isn't really enough information to determine this exactly, but you could say something like: "The proteins Celebi, Shayman, and Azelf have 6-letter names and are considered small; the proteins Artikuno, Terrakion, and Cressela have 13-letter names and are considered large."
• The "penguin / polar bear" terminology is amusing but I would go with something more informative. I'd be inclined to call the polar bears "verbs" and the penguins "suffixes".
• In fact, I'd lean right in to the analogy with sentences. "Each instruction, which I shall call a sentence, consists of some number of words separated by the space codon CND. There is also a space at the beginning and end of each sentence. The words are of two kinds: nouns, each naming a protein, and verbs, each specifying an action and 'inflected' in a manner I shall describe shortly. Sentences are in SOV order: first the subject, the protein doing the work; then the object, the protein being acted upon; then one or more verbs, indicating what is to be done. Each protein is either large (with a name 13 letters long) or small (with a name 6 letters long). Each verb has a 10-letter stem, determining what action it specifies, followed by one or two 3-letter suffixes each of which is either NAN (indicating a small protein) or NNN (indicating a large one). For the 'construct' verb, there are two of these suffixes, indicating first the size of the subject and then that of the object. For 'pack' there is just one, indicating the size of the subject. For 'cut' there is just one, indicating the size of the object. Finally, the way in which the set of actions to be performed is represented in the sentence depends on the size of the subject protein. For a small subject, the V portion of the sentence contains the verbs for the action to be performed; but for a large subject it contains the verbs for the actions not to be performed."
• Oh, I see an actual omission in your answer. What about the order in which verbs appear? For short subjects, it's the same order as in the instruction, and in the examples we have this order is always consistent: construct, cut, pack. Maybe that's obvious enough not to need mentioning. But for long subjects, the instruction doesn't explicitly specify an order for the actions not listed, but the output is always consistent with the same order as above. (I guess the idea is that these three actions only actually make sense in that order, or something.)
• I think I would want to say explicitly that despite the use of the word "codon", the letters are not consistently grouped into threes as in earth's DNA.
• It seems interesting and possibly noteworthy that the "content" is all conveyed using the letters CGAT, and the letters N,D are used only in the "space" codon and the size-marker suffixes. (C,G,A,T are the letters used to name bases in earth's DNA, though presumably we are not supposed to assume that the aliens' C,G,A,T are the same as ours.)
• You haven't said anything about how you reached the conclusions you did. I don't know whether this is a thing you are expected to do in linguistics olympiads -- it's not like mathematics where there's a clear distinction between a statement and its proof, and the proof is often more the point than the statement. But I would at least consider laying out a bit of a pathway showing how one might reach your conclusions. "The first thing we notice is that every output both starts and ends with CND -- and then that this also occurs within the output sequences. It seems plausible that this acts as a delimiter. Dividing each output sequence up at the occurrences of CND, we then see that there are six "words" that occur only in the first two sections of a sequence, and others that occur only after that." And so on. Try to show how the interpretation you ended up with is the only reasonable one, or at least the most plausible or simplest or something.

I don't know whether you've looked to see whether there's any structure within the nouns and the verb stems. I looked only very quickly and didn't see anything, and we have so few to look at that it might be hard to tell even if there is. But if I were solving this for competition purposes, I'd want to spend a few minutes staring at the words just in case there's something worth mentioning.