Bioinformatics Systems Lecture 1

Bioinformatics, also known as computational molecular biology, is an interdisciplinary field combining computer science and biological science
It uses computing to store, retrieve, manipulate, and distribute information related to biological macromolecules such as DNA, RNA, and proteins
Genomic data analysis often involves highly repetitive or complex mathematical tasks, making computers vital for this process
The question of how much biology a computer scientist needs to understand to work in bioinformatics is answered as "enough to deeply understand the biological problem and to turn it into an adequate computational problem"
The course provides a brief introduction to biological concepts
It aims to teach students how to design algorithms for solving biological problems

Essential Bioinformatics by Jin Xiong (Texas A&M University, Cambridge University Press)
An Introduction to Bioinformatics Algorithms by Neil C. Jones and Pavel A. Pevzner

Bioinformatics involves the technology that uses computers to store, retrieve, manipulate, and distribute information related to biological macromolecules
The goal of bioinformatics is the use of computers for genomic data analysis
Bioinformatics algorithms are designed to solve biological problems
Understanding how these algorithms work leads to confidence in their results and allows for identification and fixing of potential algorithm weaknesses

Cells have a life cycle: birth, eating, replication, death
Cells have complex mechanical systems
They store information to replicate, collect components to make replicas (offspring)
Prokaryotic cells are unicellular organisms like bacteria
Eukaryotic cells, like humans, are multicellular

Cells come in a wide variety of types
DNA, RNA, and proteins are the main molecules
DNA (deoxyribonucleic acid): Holds genetic instructions for cell functions
RNA (ribonucleic acid): Transfers genetic instructions and aids in protein synthesis
Proteins: Perform biochemical reactions, signal to other cells, and form body components

The nucleus is a membrane-enclosed organelle in most cells
It contains most of the cell's genetic materials organized as chromosomes
Chromosomes are long, continuous pieces of DNA that carry genes, regulatory elements, and other nucleotide sequences
Variations in chromosome shape and number differ between organisms

DNA is a long polymer chain of nucleotides
Nucleotides have four types: adenine (A), cytosine (C), guanine (G), and thymine (T)
DNA typically exists in a double helix form
Strands are complementary (A with T, G with C), making the structure readily replicable.

DNA is written in a four-letter alphabet, proteins in a 20-letter alphabet
Two types of cells: eukaryotic (with nuclei) and prokaryotic (without nuclei)
In eukaryotes, DNA resides in the nucleus, whereas protein synthesis happens outside the nucleus (in cytoplasm)
Messenger RNA (mRNA) copies DNA segments (genes) and takes the genetic information to ribosomes
Ribosomes synthesize proteins based on the mRNA sequence
Chemically, RNA is similar to DNA but has uracil (U) instead of thymine (T)
RNA is typically single-stranded, allowing it to be active in cellular processes

Genes are segments of DNA that hold hereditary information
Genes are transcribed into RNA
RNA is translated into proteins (amino acid sequences)
This process (transcription and translation) is called gene expression
The flow of this information (DNA -> RNA -> Protein) is central to molecular biology

Databases store biological data: primary, secondary, and specialized
Primary databases contain original data (GenBank, Protein Data Bank (PDB))
Secondary databases analyze original data and contain processed information, often with annotations (e.g. SWISS-Prot)
Specialized databases focus on specific organisms or data types (e.g. Flybase, HIV sequence database.

An algorithm is a sequence of instructions to solve a problem
Pseudocode describes algorithms in a way that is not too formal or too specific
A problem describes a class of computational tasks, and one instance of a problem is one particular input.

Algorithms should correctly translate inputs into outputs for every possible input
Fast/Slow Algorithms: time complexity/efficiency is determined by the number of operations

Computer scientists often classify algorithms in categories based on the way their time complexity relates to input size.
The fastest growing term dictates the classification.
Big-O notation focuses on how the growth rate behaves as the input size gets larger.
For analysis purposes, the constant multipliers are generally ignored.

Several algorithm design techniques exist, like:
Exhaustive search (brute force): A general approach that checks all possible solutions
Branch-and-Bound: Optimization techniques allowing for eliminating inappropriate search branches, speeding up the process
Greedy algorithm: Aims to make the best local choice, leading to the eventual solution, possibly not optimal.
Dynamic Programming: optimization technique to divide problems into subproblems repeatedly, solving them only once and saving time by avoiding repeated computations
Divide-and-Conquer: An approach to break a larger problem into smaller, independent subproblems, solve them and then recombine those solutions to produce a solution for the initial larger problem
Machine Learning algorithms use statistical analyses of previously collected data to produce results.
Randomized algorithms involve using randomness as part of the basic strategy.