Bioinformatics Intro
What is Bioinformatics
Bioinformatics is the use of computational methods to store, analyze, interpret, and model biological data.
Bioinformatics often focuses on biological data analysis, while computational biology often focuses more on modeling biological systems.
Bioinformatics Subfields
| Subfield | Focus | Applications |
|---|---|---|
| Genomics (Omics Data Analysis#What are Omics) | sequencing, assembling, analyzing genome | genome sequencing, genetic variation studies, gene expression analysis |
| Transcriptomics (Omics Data Analysis#What are Omics) | RNA transcripts produced by the genome | gene expression profiling, disease diagnosis, understanding cell behavior |
| Proteomics (Omics Data Analysis#What are Omics) | protein structures, functions, and interactions | protein identification in diseases, biomarker discovery, drug-target identification |
| Metabolomics (Omics Data Analysis#What are Omics) | metabolites and metabolic processes | biomarker discovery, disease metabolism analysis, personalized medicine |
| Structural bioinformatics | 3D structure of biological macromolecules (e.g. proteins, nucleic acids) | protein folding analysis, drug design, enzyme mechanism studies |
| Pharmacogenomics | drug-gene interactions | tailoring drug prescriptions, improving therapeutic efficacy, reducing side effects |
| Systems biology | modeling of biological networks and systems | understanding cellular networks, modeling disease mechanisms, gene-regulatory network analysis |
| Cheminformatics | chemical data for biological applications | drug discovery, analyzing chemical libraries, predicting compound effects |
| Epigenomics | chemical modifications in gene expression | cancer gender regulation, aging studies, understanding inheritance patterns |
Bioinformatics Databases
| Database Type | Definition | Key Features | Examples |
|---|---|---|---|
| primary | repositories that contains original biological data | raw data; frequently updated | GenBank/NCBI, EMBL-EBI, DDBJ |
| secondary | data compiled from primary databases: adding curated, value-added information like annotations | curation and annotation of data; provides additional context and insights | UniProt, Pfam, RefSeq |
| derived / integrated | data derived from primary and secondary databases through computational analysis | integration of data from multiple sources; provides comprehensive insights and functional annotations | Ensembl, Gene Ontology (GO), InterPro |
Sequence Analysis
Pairwise Sequence Alignment
= Aligning two biological sequences (DNA, RNA, or protein) to find similarity, differences, and evolutionary relationships.
Main types
- Global alignment: align entire sequences end-to-end
- Local alignment: find best matching subsequences
Key components
- Scoring system
- Match (e.g., +1)
- Mismatch (e.g., −1)
- Gap penalty (penalty for insertions/deletions)
- Gap penalties
- Linear: same cost per gap
- Affine gap penalty (more realistic):
- Gap penalty=gap opening+(gap length×gap extension)
- Gap opening penalty: high
- Gap extension penalty: lower
Tips
A gap is an artificial placeholder (-) inserted into a sequence so that two sequences can be aligned properly. It represents a biological event: Insertion (something added in one sequence) or Deletion (something lost in the other).
Methods
| Needleman–Wunsch | Smith–Waterman | BLAST | |
|---|---|---|---|
| Type | Global alignment | Local alignment | Local (heuristic) alignment |
| Accuracy | Exact | Exact | Approximate |
| Speed | Slow | Very slow | Fast |
| Use case | Full sequence comparison | Motifs/domains | Database search |
| Gap handling | Often affine | Often affine | Simplified |
| Tools | EMBOSS Needle | EMBOSS Smith | BLAST (NCBI BLAST) |
Limitations and challenges
- algorithm complexity
- scoring system selection: Alignment quality depends heavily on the scoring scheme
- gap penalties: Difficult to correctly model gaps representing biological insertions/deletions
- sequence length: Sequence length affects both computation and biological interpretability
Multiple Sequence Alignment
= Aligning 3 or more biological sequences (DNA, RNA, or proteins) simultaneously to identify:
- Conserved regions
- Functional motifs
- Evolutionary relationships
Main types
- Global alignment: align entire sequences end-to-end
- Local/Motif-focused alignment: find best matching subsequences
Methods
| Progressive | Iterative | Consistency-based | HMM/Profile | |
|---|---|---|---|---|
| How it works | build pairwise distances -> align sequences step-by-step | start with initial alignment-> repeatedly refine it | combines multiple pairwise alignments | use probabilistic models |
| Accuracy | Medium | High | Very high | Very high (for families) |
| Speed | Fast | Medium | Slow | Medium |
| Use case | General-purpose MSA | Higher accuracy alignment | Difficult/divergent sequences | Protein families, remote homology |
| Tools | Clustal Omega, ClustalW | MUSCLE, MAFFT | T-Coffee | HMMER |
Limitations and challenges
- computational complexity
- erorr propagation
- sensitivity to parameters
Structural Alignment
= Aligning based on the 3D structure of biological molecules
Homology Searching
= Finding sequences in a database that are evolutionarily related to a query sequence
Important distinction
- Similarity = measurable (alignment score)
- Homology = biological conclusion (shared ancestry)