Leucine, arginine, tryptophan. There are, of course, other solutions.
ACCGAGAGCTTGAAAGGCAGCTAGTGCCATTTGTT
The two pieces left after the cut are:
ATGG
TACCTTAA
AATTCCGCATCCTTCACTG
GGCGTAGGAAGTGAC
which have 12 and 34 nucleotides, respectively.
One possible solution is:
Three colors suffice. One possible solution is:
There are three optimal alignments:
AT-TC
ATGGA
ATT-C
ATGGA
ATTC-
ATGGA
Applying the basic algorithm and recovering the optimal alignments from the matrix we can see that no other optimal alignments exist.
To find βΔt, we equate 8/80 to the formula for Pr(transversion) given in the question. The result will be (1/4)*ln(1.25) or approximately 0.055786.
To find αΔt, we equate 5/80 to the formula for Pr(transition) given in the question, and use the previous value found for βΔt. The result will be (1/4)*ln(0.2/(0.3875)^2) or approximately 0.071660.
0.2 for each amino acid.
3-letter code ok.
1-letter code 0.2 for all.
nucleotides, enzymes, etc.: 0.0
just reverse or just complement: 0.2
error in base: -0.1
wrong pairing: 0.0
1 strand only, right cut: 0.5
1 strand only, wrong cut: 0.2
2 strands, right cut, sum ok: 1.0
2 strands, right cut, no sum: 0.8
2 strands, wrong cut: 0.6
0.5 value
1.0 value
Hamiltonian path: 0.8
2 cycles instead of one: 0.6
just one cycle, not Hamiltonian: 0.2
0.5 value
edge coloring: 0.1
cover by paths: 0.0
alignments but no reasonable explanation: 0.5
only one alignment: 0.3
βΔt: 2.0
αΔt: 3.0
went halfway through: half credit
used count instead of frequency: 1.5
plugged value in wrong place: 1,0
just last calculation step wrong: -0.5
© 2015 Joao Meidanis