Let \(P_w(n)\) be the set of distinct subwords (intervals) in a word \(w\).
Let \(p_w(n)\) be the cardinality of \(P_w(n)\).
Let \(f(c)\) be the sequence in FASTA with 4-symbol Protein Data Bank code \(c\).
\(|P_{f(5GLN_1)}(2) \setminus P_{f(3EIM_1)}(2)|=91\),
\(|P_{f(3EIM_1)}(2) \setminus P_{f(5GLN_1)}(2)|=68\).
Let
\(
Z_k(x,y)=|P_x(k)\setminus P_y(k)|+|P_y(k)\setminus P_x(k)|
\)
be a LZ76 style (set of subwords) Jaccard distance numerator for \(P(k)\).Hydrophobic-polar version of Sequence 1:11000000000011110100101110010010101011010101010001011010001100101000011010011101000111101001111100111101100010001011100000110111110001111101000110100010111100001000101111111010010010011001000010000111111011110001101101100101100010111110000011011110000100010000100001101010010111000111101111100000110101010110000010110001001010010001010100101010
Let d be the
Otu--Sayood
distance d.
Let d1 be the Otu--Sayood distance d1. (This makes the 4TYN sequence AAAAAA a close match...)
A roughly speaking expected distance is \((0.85)(0.8)(\frac{660
}{\log_{20}
660}-\frac{316}{\log_{20}316})=95.2\)
Status
Protein1
Protein2
d
d1/2
Query variables
5GLN_1
3EIM_1
122
115
Was not able to put for d Was not able to put for d1
In notation analogous to
[Theorem 16, Kjos-Hanssen, Niraula and Yoon (2022)],
\[
\delta=
\alpha \mathrm{min} + (1-\alpha) \mathrm{max}=
\begin{cases}
d &\alpha=0,\\
d_1/2 &\alpha=1/2
\end{cases}
\]