« return to puzzle

Memo From Agent C

Gavin Schnitzler

The second and fourth entries before the break (---) are straight-out DNA sequences. To get instructions for the rest of the puzzle, "transcribe" these DNAs into RNA (everything the same except T is replaced by U). Next "translate" the RNA into amino acids (each 3 base pair codon encodes one amino acid, and the same encoding is used by all living things). Do this for all three reading frames (e.g. starting with the first, second or third base) and then reduce the amino acid to its single letter code. For instance ATG= Methionine or M, GGG=Glycine or G. This can be done by hand using a codon table or easily online at sites like: http://www.vivo.colostate.edu/molkit/translate/index.html.

For the second entry frame 1 starting from the first base gives garbage except for the words "Red herring". To get the REAL message you must follow the rules for protein translation from messenger RNA (mRNA). Always start at ATG (Met=M) and stop at an in frame UAA, UGA or UAG stop codon. This gives a series of linked amino acids that form a polypeptide or protein "clue". The correct way to interpret each entry is also hinted at by the question revealed starting at base 1 in message 4 (START WHERE?). For biology sticklers, an ATG is started at best if it is preceded by (A/G)CC in eukaryotes (the Kozak sequence), or by AGGAGG(N4-7) in prokaryotes (the Shine-Dalgarno sequence). These sequences are present in the clues, but you can get the correct answer just by using the first ATG encountered.

Doing this gives this message from frame 2 of second entry "MAY DECRYPT KEEPING ACGT SPACING CYCLICALLY THEN TRANSLATE MRNAS". This means that all the other entries can be converted to DNA sequence by fitting them to an alphabet (circularized so that Z wraps back to A) in such a way that the distance between letters is the same as A – gap of 1- C- gap of 3 – G – gap of 12 – T. For instance entry 1 before the break has only WJSQ. Fitting to the ACGT spacing, Q – 1- S – 3 – W – 12 – J, means Q=A, S=C, W=G & J=T. Use this trick to convert all entries to DNA sequences. Then convert to RNA and translate as above.

Doing so for all entries before the break and translating the amino acid sequence properly gives you three new instructions to use in solving all entries after the break:

MIND THAT ESCHERICHIA YIELDED FIRST AND THIRD
Escherichia Coli is a common human gut bacterium and lab strain. Bacteria can have polycistronic mRNAs that can encode more than one protein. If there is another ATG after a stop codon (especially if following a Shine-Dalgarno sequence), a second protein is made. Hence, look for more than one clue in the first and third entries, e.g., line 1 gives the two clues MAKETENTIMESTENTIMESTEN and MYSELF.

MAMMALIAN GENES GAVE ALL THE REST
*Note that in various frames you can also get the words, WEASEL PIG CAT CHIMP & MAN, but since these don’t appear when the mRNA is properly translated they are simply distractors.

Almost all mammalian mRNAs are "spliced," meaning that after the DNA is copied to make RNA (in the process of transcription), parts of the RNA sequence are excised to give a shorter final mRNA. The rules for splicing are as follows: when you see the sequence: GGT(A or G)AG, followed by any length of intron/filler DNA, followed by an ~13-20 base stretch that is rich in T’s and C’s (any base)(C or T)AG, remove all that DNA except for the first G and hook it up to the remaining DNA sequence. This can be looked up readily in any Molecular Bio text or online.

*Note that splicing is now clued in several ways that were not in the original puzzle:

This will yield a very different amino acid sequence and clue from the unspliced mRNA. For instance:

ATGGCTAAAGAAACTGCTCTTAAAATTAATGGTGAGAATGGTGCTGGTGAAATGGAAAA
TACTTGGATTACTCATGCTGATATTTCTACTGCTAATACTTTTCGTATAG
AAAATGATTGA...

reads MAKE TALKING ENGAGEMENT WITH A DISTANT FRIEND.

Properly processed (removing all sequence in italics) it gives ATGGCTAAAGAAACTGCTCTTAAAATTAATGAAAATGATTGA… which reads MAKE TALKIN’ END.

Now you want to consider the last instruction entry before the break, which read "MINIMAL ANSWERS TRANSLATED AS NECESSARY EACH ALIGNED VIA ADENINES PLACE"

The first part of "Minimal answers" indicates that for each polypeptide clue you must derive a one letter answer. This might be clued directly, or it might be clued by giving a 3 letter codon that can be translated into a single letter.

The second part "Each aligned via adenines place" means that you line up these single letters in alphabetical order based on the letter used to represent adenine or A in the original encoded message (e.g. for the 1st entry before the break A was represented as Q).

The clues, answers and letter used to represent A are as follows, in order:

  1. A=B, MAKE TEN TIMES TEN TIMES TEN (M) + MYSELF (I). Polycistronic: therefore two clues and two one-letter answers.
  2. A=C, MERGE THIS MELANGE TRIPLET WITH D YIELDING AIRPIPE. "Duct" can be made from the "mélange" (part DNA part RNA) triplet UCT + D. UCT is the answer. Translated (as UCU) to an amino acid gives "S."
    *Note if you don’t splice properly, you get MERGE THIS MELANGE TRIPLET WITH D YIELDING EVIDENTIAL FINDINGS, which would give ATA form DATA or "I".
    **This one is a bit kludgy, but I couldn’t think of a good way to clue "S" and I like the misdirection with DATA.
  3. A=G, MERRY WIVES PART (ACT = T) + MEANING GREAT GRADE (A). Polycistronic: therefore two clues and two one-letter answers.
  4. A=M, MANY A CAR DRIVERS SERVICE AID (AAA=K)
    * If not properly spliced leads to an any early stop codon and gives the shorter misleading clue MANY A CAR.
  5. A=T, MAKE TALKIN’ END (GAG = E)
    * Improperly spliced gives the misleading MAKE TALKING ENGAGEMENT WITH A DISTANT FRIEND
    ** The contraction on "talkin’" is unfortunate, but it was the only way I could get the splice concensus in.
  6. A=Y, MEANING ELEMENT SEVEN (N, for Nitrogen).
    * If not properly spliced gives the misleading answer MEANING ELEMENTAL SILVER (Ag).

Reading all single letters in order gives the final answer MISTAKEN.

Hence, going back to the flavor text at the beginning: "the murderer’s identity" is MISTAKEN… or it’s a case of mistaken identity.

*Note that the first M could also be G or K, and the I could be "C" for Agent C’s name. These are the only two letters that are ambiguous, however, and in context the only word that can be made uses M and I.

Note that some English in the clues is a bit strained. This could not be completely avoided, because there are no amino acids represented by the letters B, J, O, U, X or Z. Avoiding any words containing O or U was particularly annoying. ;)