Key Concepts in Bioinformatics and Protein Structure

Hidden Markov Model (HMM)

A statistical model used for sequence analysis; can model protein domains from multiple sequence alignments.

Position Weight Matrix (PWM)

Represents the frequency of each nucleotide or amino acid at each position in a sequence motif; used in motif discovery.

Inferential Statistics

Statistical testing to determine if gene expression differences between conditions are significant (e.g., t-tests, FDR correction).

Normalization (RNA-seq)

Adjusts RNA-seq read counts for sequencing depth and gene/transcript length (e.g., TPM, RPKM, DESeq2).

Clustering Analysis

Groups genes/samples with similar expression patterns; does not imply causality or statistical significance.

Differentially Expressed Genes (DEGs)

Genes with statistically significant differences in expression between conditions. Identified using inferential stats.

RNA-Seq

Sequencing method for transcriptome profiling; detects novel transcripts and isoforms.

DNA Microarray

Chip-based expression analysis; limited to known genes/transcripts.

DAVID

Identifies overrepresented functional terms (GO, KEGG) in gene lists.

GEO2R

Online tool to find differentially expressed genes from GEO datasets.

MEME

Discovers sequence motifs (e.g., TF binding sites) in DNA/protein sequences.

Cufflinks

Assembles transcripts and quantifies RNA-seq data; not for functional enrichment.

Domain

A structurally/functionally independent unit in a protein.

Fold

The core 3D structure of a protein domain.

Secondary Structure

Local structures (alpha helices, beta sheets), stabilized by hydrogen bonds.

Tertiary Structure

The complete 3D shape of a single polypeptide.

Quaternary Structure

The 3D arrangement of multiple polypeptides in a complex.

Intrinsically Disordered Proteins (IDPs)

Proteins/regions lacking fixed 3D structure; still functional.

RCSB Protein Data Bank (PDB)

Database of experimentally-determined 3D structures of proteins/nucleic acids.

VAST (Vector Alignment Search Tool)

Compares 3D protein structures to find structural similarities.

CATH Database

Classifies protein structures by: Class, Architecture, Topology, and Homologous superfamily.

DALI

Compares 3D protein structures via inter-residue distance matrices.

AlphaFold

DeepMind's AI tool for accurate 3D protein structure prediction.

Molecular Docking

Predicts molecular interactions (e.g., ligand-protein binding); useful in drug discovery.

Homology Modeling

Predicts protein structure using a known template with sequence similarity.

Ab initio Prediction

Predicts protein structure from scratch, without templates.

Artificial Neural Network

Used for secondary structure prediction (e.g., PSIPRED).

Phylogenetic Tree

Diagram showing evolutionary relationships.

Neighbor-Joining

Distance-based method for tree building.

Maximum Likelihood

Builds tree based on most probable evolutionary model.

Maximum Parsimony

Builds tree with the fewest evolutionary changes.

k-means Clustering

Groups data using ML; not used for phylogenetic trees.

Scale-Free Network

Network with hub nodes; follows power-law distribution.

Small-World Property

Most nodes connected through a small number of steps; common in biological systems.

Random Network

Edges randomly distributed; lacks structure of biological networks.