Scribd
Upload
40 views

What Are Some Open Problems in Bioinformatics

The document discusses several open problems in bioinformatics including the molecule isomorphism problem, string search algorithms, pairwise sequence alignment algorithms, protein and RNA folding problems, and biological network inference. Many of these problems relate to determining isomorphisms or structures given molecular representations and sequences.

Uploaded by

murat
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
40 views

What Are Some Open Problems in Bioinformatics

The document discusses several open problems in bioinformatics including the molecule isomorphism problem, string search algorithms, pairwise sequence alignment algorithms, protein and RNA folding problems, and biological network inference. Many of these problems relate to determining isomorphisms or structures given molecular representations and sequences.

Uploaded by

murat
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

What are some open problems in bioinformatics?

Some open theoretical and practical problems I have encountered off the top of my head
:

• Molecule Isomorphism Problem, given two representations (be it


sequences, structures, notations etc) of molecule R1R1 and R2R2, infer the
isomorphism of the two. Could it be solved in deterministic polynomial
time? A variation of graph isomorphism problem. For highly reduced
form of canonical molecules like antibody and antibody drugs conjugate
(ADC) my own constructed algorithms seem to always fall in polynomial
time. For more complex molecules, however, the problem approaches
the general problem in computer science complexity-wise. As such no
definite answer is known until P=NPP=NP is solved.
• String search algorithms. What is the lower bound of string search
algorithms? Applying linear space constraints, what is the lower bound of
string search algorithms?
• Pairwise sequence alignment algorithms. What is the lower bound of
pairwise sequence alignment algorithms? Applying linear space
constrains, what is the lower bound of pairwise sequence alignment
algorithm?
• Molecule database search. Does error-free database system for
molecules exist? At the end this goes back to first problem (Molecule
Isomorphism Problem), an existing general algorithm for graph
isomorphism in polynomial time means that it's possible to have 100%
accuracy in molecular search, time and space (polynomially)
unconstrained.
• Protein and RNA Folding Problem, given the
sequence s0s1s3…sns0s1s3…sn of a protein or RNA, infer the secondary,
tertiary and quartenary structure of it in a given environment EvEv. Does
such algorithm exist? Is it beyond PP?
• Reverse Protein and RNA Folding Problem. Given a representation (be
it distance matrix, coordinate vector, vector graphic etc) of protein or RNA
structure RsRs, find an underlying sequence SS that would adopt the
structure. Does such algorithm exist? If so, is it beyond PP?
• Dynamic Substructure Prediction. Given the structure RsRs of a protein,
applying
change Ti:(s0s1…sn)→(s0s1…s′0s′1…s′m…sn)Ti:(s0s1…sn)→(s0s1…s0′s1′…
sm′…sn) (insertion), Td:(s0s1…sn)→(s0s1…si−1sisjsj+1…sn)Td:(s0s1…sn)→(
s0s1…si−1sisjsj+1…sn) where j−i>1j−i>1 (deletion),
or Ts:(s0s1…sk…sn)→(s0s1…s′k…sn)Ts:(s0s1…sk…sn)→(s0s1…sk′…sn) wh
ere sk≠s′ksk≠sk′ (substitution) to its underlying sequence, infer the new
structure R′sRs′. Does an algorithm exist for all non-secondary parts of
protein (turn and coil)? Is it beyond PP?
• De novo protein design. Is it possible to design a protein from scratch?
Does such algorithm/method exist?
• Mathematical biological network inference. Is it possible to entirely
infer biological networks mathematically (i.e. without using database
search)? Is it possible to mathematically infer causality among biological
entities? Is it possible to mathematically predict pathways?

You might also like