Biological Foundations for Bioinformatics

The Central Dogma of Molecular Biology

Central Dogma: DNA to RNA to Protein

The central dogma describes information flow in biological systems: DNA is transcribed into RNA, which is translated into protein. This unidirectional flow (with known exceptions like reverse transcription) underpins all of computational biology.

DNA Structure and Function

DNA is a double-stranded helical polymer of deoxyribonucleotides. Each nucleotide consists of a deoxyribose sugar, a phosphate group, and one of four nitrogenous bases: adenine (A), thymine (T), guanine (G), or cytosine (C). Watson-Crick base pairing (A-T with two hydrogen bonds, G-C with three) creates the antiparallel double helix. The 5' to 3' directionality of each strand is critical for replication and transcription.

Replication: Semi-conservative, catalyzed by DNA polymerase III (prokaryotes) or polymerases alpha/delta/epsilon (eukaryotes)
Supercoiling: Managed by topoisomerases; relevant to genome topology and chromatin organization
Telomeres: Repetitive sequences (TTAGGG in humans) protecting chromosome ends, maintained by telomerase

RNA Biology

RNA uses ribose and substitutes uracil (U) for thymine. Key RNA species:

mRNA: Carries coding information; undergoes 5' capping, polyadenylation, and splicing in eukaryotes
tRNA: Adapter molecules with anticodon loops; aminoacyl-tRNA synthetases charge them with cognate amino acids
rRNA: Structural and catalytic component of ribosomes (16S/18S for small subunit, 23S/28S for large)
Non-coding RNA: miRNA (~22 nt, post-transcriptional silencing), lncRNA (>200 nt, diverse regulatory roles), siRNA, piRNA, snoRNA, circRNA

Proteins

Proteins are polymers of 20 standard amino acids linked by peptide bonds. Each amino acid has an amino group, carboxyl group, and a variable R-group determining its physicochemical properties (hydrophobic, polar, charged, aromatic). Protein function depends on three-dimensional structure, which is determined by primary sequence through thermodynamic folding.

The Genetic Code

The genetic code maps 64 codons (triplets of nucleotides) to 20 amino acids plus three stop signals. Key properties:

| Property | Description | |---|---| | Degeneracy | Most amino acids encoded by multiple codons (wobble position) | | Universality | Nearly universal across life; exceptions in mitochondria, Mycoplasma, ciliates | | Non-overlapping | Codons read sequentially without overlap in standard translation | | Unambiguous | Each codon specifies exactly one amino acid |

Codon usage bias: Organisms preferentially use certain synonymous codons, correlated with tRNA abundance. The Codon Adaptation Index (CAI) quantifies this bias and predicts expression levels.

Gene Expression and Regulation

Transcription

Prokaryotic transcription uses a single RNA polymerase with sigma factors for promoter recognition (-10 and -35 elements). Eukaryotic transcription employs three polymerases (Pol I: rRNA, Pol II: mRNA, Pol III: tRNA/5S rRNA) with complex promoter architecture (TATA box, CpG islands, enhancers, silencers).

Post-Transcriptional Processing

Splicing: Intron removal by the spliceosome (U1, U2, U4, U5, U6 snRNPs); alternative splicing generates proteome diversity (>95% of multi-exon human genes)
RNA editing: Adenosine-to-inosine (ADAR) and cytidine-to-uridine deamination

Translation

Ribosomes decode mRNA in three phases: initiation (Shine-Dalgarno/Kozak sequence recognition), elongation (codon-anticodon matching, peptide bond formation), termination (release factors at stop codons). Post-translational modifications (phosphorylation, glycosylation, ubiquitination) expand functional diversity.

Genome Organization

Prokaryotic Genomes

Typically single circular chromosomes (1-10 Mb), gene-dense (~85-95% coding), organized into operons with polycistronic mRNA. Horizontal gene transfer via conjugation, transduction, and transformation drives genomic plasticity. Plasmids carry accessory genes (antibiotic resistance, virulence factors).

Eukaryotic Genomes

Linear chromosomes with centromeres and telomeres. The human genome (~3.2 Gb) is only ~1.5% protein-coding. Major non-coding components:

Repetitive elements: SINEs (Alu, ~11%), LINEs (L1, ~17%), DNA transposons (~3%), LTR retrotransposons (~8%)
Regulatory sequences: Enhancers, silencers, insulators, locus control regions
Structural elements: Centromeric satellite DNA, heterochromatin

Chromatin and Epigenetics

DNA wraps around histone octamers (H2A, H2B, H3, H4) forming nucleosomes. Epigenetic modifications regulate gene expression without altering DNA sequence:

DNA methylation: CpG dinucleotides; 5-methylcytosine associated with silencing; maintained by DNMT1, established by DNMT3a/b
Histone modifications: Acetylation (activation, HATs/HDACs), methylation (context-dependent, e.g., H3K4me3 = active, H3K27me3 = repressive), phosphorylation, ubiquitination
Chromatin remodeling: ATP-dependent complexes (SWI/SNF, ISWI, CHD, INO80) reposition nucleosomes

Mutations and Variation

Mutation Types

Point mutations: Transitions (purine-purine or pyrimidine-pyrimidine), transversions (purine-pyrimidine)
Insertions/deletions (indels): Frameshifts if not multiples of 3 in coding regions
Structural variants: Inversions, translocations, duplications, copy number variations (CNVs)
Microsatellite instability: Expansion/contraction of short tandem repeats (STRs)

Functional Impact

Variants are classified by consequence: synonymous, missense (conservative vs. radical), nonsense, splice-site, regulatory. Pathogenicity prediction tools (SIFT, PolyPhen-2, CADD, REVEL) integrate evolutionary conservation and structural features.

Biological Databases

Primary Sequence Databases

GenBank/ENA/DDBJ: INSDC partnership; nucleotide sequences with annotations; >2 billion sequences
UniProt: Protein sequences; UniProtKB/Swiss-Prot (curated, ~570K entries) and TrEMBL (automated, >200M entries)
RefSeq: NCBI reference sequences with non-redundant, curated records

Structure Databases

PDB (Protein Data Bank): 3D macromolecular structures from X-ray crystallography, cryo-EM, NMR; >200K structures
AlphaFold DB: Predicted structures for >200M proteins
SCOP/CATH: Hierarchical protein structure classification

Functional and Pathway Databases

Gene Ontology (GO): Controlled vocabulary for molecular function, biological process, cellular component
KEGG: Metabolic pathways, signaling networks, disease associations
Pfam/InterPro: Protein domain families and functional annotations
OMIM: Mendelian disease-gene relationships

Bioinformatics File Formats

Sequence Formats

FASTA: Header line starting with > followed by identifier, then sequence on subsequent lines. Simple, universal, no quality information.

>sp|P00533|EGFR_HUMAN Epidermal growth factor receptor
MRPSGTAGAALLALLAALCPASRALEEKKVCQGTSNKLTQLGTFEDHFLSLQRMFNNCEVV

FASTQ: Extends FASTA with per-base quality scores (Phred+33 encoding). Four lines per read: header (@), sequence, separator (+), quality string. Quality score Q = -10 log10(P_error); Q30 means 1 in 1000 error probability.

Alignment Formats

SAM/BAM: Sequence Alignment/Map format. SAM is tab-delimited text; BAM is the binary compressed equivalent. Fields include QNAME, FLAG (bitwise: mapped, paired, strand), RNAME, POS, MAPQ, CIGAR string (alignment operations: M/I/D/S/H/N), mate information, sequence, quality. The CIGAR string encodes the alignment compactly (e.g., 50M2I30M1D20M).

CRAM: Reference-based compression of alignment data; more space-efficient than BAM.

Variant Formats

VCF (Variant Call Format): Stores genomic variants. Header lines (##) specify metadata, INFO/FORMAT field definitions, and reference genome. Data lines contain CHROM, POS, ID, REF, ALT, QUAL, FILTER, INFO, and per-sample genotype fields. Multi-allelic sites, structural variants, and phasing information are supported.

BED: Tab-delimited genomic intervals (chrom, start, end) with optional name, score, strand fields. Zero-based, half-open coordinate system.

Annotation Formats

GFF3/GTF: Gene annotations with nine columns (seqid, source, type, start, end, score, strand, phase, attributes). GTF uses a structured attribute format; GFF3 uses key=value pairs with hierarchical parent-child relationships.

Computational Considerations

Biological data exhibits particular computational properties: sequences are strings over small alphabets (4 for DNA, 20 for protein), genome-scale analyses involve gigabytes to terabytes of data, and biological signal is often subtle against a background of evolutionary noise. Understanding these biological foundations is essential for correct algorithmic design and meaningful interpretation of computational results.