Conceptual Foundation
Biological observations originate from experiments, computational predictions, clinical measurements, field studies, and scientific databases. Computers cannot interpret biological meaning directly from an unstructured observation; the information must first be represented in a suitable digital form. For example, DNA can be represented as a sequence of nucleotide symbols, gene expression as a numerical matrix, and protein structures as collections of three-dimensional atomic coordinates. A correct computational representation preserves essential biological meaning while allowing efficient analysis.
Major Biological Data Types
| Data Type | Description | Common Symbols | Computational Representation | Scientific Uses | Important Considerations |
|---|---|---|---|---|---|
Nucleotide Sequence Data | Represents DNA or RNA as an ordered sequence of nucleotide symbols. |
| A character string or sequence object, commonly stored in formats such as FASTA or FASTQ. |
| Sequence orientation, ambiguous nucleotide symbols, sequence quality, organism identity, and reference coordinates must be interpreted correctly. |
Protein Sequence Data | Represents a protein as an ordered chain of amino-acid residues. | A sequence of one-letter amino-acid codes, often stored in FASTA format. |
| Amino-acid sequences do not directly provide a complete description of protein conformation, post-translational modifications, or biological activity. | |
Gene Expression Data | Represents the measured abundance of transcripts or other expression-related quantities across genes, samples, or experimental conditions. | A numerical matrix in which rows may represent genes and columns may represent biological samples. |
| The numerical values may represent raw counts, normalized counts, transformed expression values, or other measurements. These representations are not interchangeable. | |
Genomic Variant Data | Represents differences between an observed genome and a specified reference sequence or another defined comparison. | Variant records containing genomic location, reference allele, alternate allele, and associated annotations. |
| Genome assembly version, coordinate convention, reference allele, alternate allele, and sample metadata must be recorded. | |
Protein Structural Data | Represents the three-dimensional organization of proteins and other biological macromolecules. | Atomic coordinates, residue identifiers, chain information, and associated structural annotations. |
| Coordinate systems, missing residues, alternate conformations, experimental resolution, and model confidence influence interpretation. | |
Molecular Interaction Data | Represents relationships among proteins, genes, metabolites, regulatory elements, or other molecular components. | Tables, graphs, networks, or structured interaction records. |
| An interaction record must be interpreted according to the type of relationship, experimental evidence, confidence, and biological context. | |
Biological Image Data | Represents observations from microscopy, histology, medical imaging, and other imaging technologies. | Pixel or voxel arrays with image metadata, spatial dimensions, and acquisition information. |
| Resolution, calibration, staining, image acquisition settings, and preprocessing affect the interpretation of measurements. | |
Clinical and Phenotypic Data | Represents observable characteristics, disease-related information, physiological measurements, and clinical observations. | Structured tables, coded clinical records, controlled vocabularies, and linked metadata. |
| Data privacy, consent, measurement definitions, missing values, population characteristics, and clinical coding systems are essential. |
Computational Representations
| Representation | Explanation | Example | Advantages | Limitations |
|---|---|---|---|---|
Strings | Used for ordered symbolic sequences such as DNA, RNA, and proteins. | ATGCGTAC |
|
|
Arrays and Matrices | Used for numerical measurements, expression profiles, and structured experimental datasets. | A gene-expression matrix with genes as rows and samples as columns. |
|
|
Tables and Records | Used to organize genomic features, variants, sample information, annotations, and experimental observations. | A table containing gene identifiers, chromosome locations, and functional descriptions. |
|
|
Graphs and Networks | Used to represent relationships among genes, proteins, metabolites, organisms, or biological pathways. | A protein interaction network in which nodes represent proteins and edges represent reported interactions. |
|
|
Three-Dimensional Coordinates | Used to describe molecular structures through spatial coordinates of atoms or other structural components. | A protein structure containing the Cartesian coordinates of its atoms. |
|
|
Images and Multidimensional Arrays | Used to represent microscopy, medical imaging, and spatial molecular measurements. | A three-dimensional microscopy image represented as a voxel array. |
|
|
Biological Data Formats
| Format | Purpose | Typical Structure | Common Use | Important Note |
|---|---|---|---|---|
FASTA | Stores nucleotide or protein sequences with descriptive header lines. | A header beginning with the greater-than symbol followed by one or more sequence lines. | Sequence storage, sequence alignment, and similarity searching. | Basic FASTA files do not normally preserve per-base sequencing quality scores. |
FASTQ | Stores nucleotide sequence reads together with corresponding per-base quality information. | Four logical lines per record: identifier, sequence, separator, and quality string. | High-throughput sequencing data processing and quality assessment. | The quality string uses an encoding scheme whose interpretation depends on the FASTQ specification and platform conventions. |
GenBank | Stores nucleotide sequence records together with biological annotations and associated metadata. | Sequence retrieval, feature annotation, and molecular research. | It can include information about genes, coding sequences, references, and other annotated features. | |
GFF3 | Represents genomic features such as genes, transcripts, exons, and other annotated regions. | Genome annotation and feature-coordinate analysis. | Coordinate conventions, feature relationships, and attribute fields must be interpreted according to the specification. | |
BED | Represents genomic intervals and associated feature information. | Genome browser tracks, interval comparison, and genomic region analysis. | BED commonly uses zero-based, half-open coordinates, which differ from some other biological formats. | |
VCF | Represents genetic variants and associated information. | Variant storage, filtering, and genomic variation analysis. | Reference genome information, variant normalization, and genotype interpretation are essential. | |
PDB and mmCIF | Represent three-dimensional macromolecular structures. | Protein structure visualization, structural comparison, and molecular modelling. | The two formats differ in organization and capabilities; mmCIF supports richer structural information and is central to the modern PDB archive. | |
SAM and BAM | Represent sequence alignment records, with BAM being a compressed binary representation of SAM. | Storage and analysis of reads aligned to reference genomes. | Alignment flags, reference coordinates, mapping quality, and optional tags have defined meanings. |
Metadata And Data Provenance
Definition
Metadata describes the context, origin, structure, and interpretation of a dataset, while provenance records how the data were generated, processed, transformed, and analyzed.
Essential Metadata
Organism or biological source.
Sample identifier and sample type.
Tissue, cell type, or environmental origin where applicable.
Experimental condition and treatment information.
Collection date or relevant time point.
Sequencing, imaging, or measurement platform.
Reference genome or database version where applicable.
Units, coordinate conventions, and measurement definitions.
Preprocessing procedures and quality-control information.
Software versions and analytical parameters.
Importance
Without appropriate metadata and provenance, researchers may be unable to reproduce an analysis, compare datasets correctly, or interpret the biological meaning of a measurement.
Data Quality And Validation
| Aspect | Description |
|---|---|
Completeness | Determine whether essential fields, sequence regions, sample measurements, and metadata are missing. |
Validity | Check whether values, identifiers, symbols, and records conform to the expected biological and technical format. |
Consistency | Check whether the same biological entity is represented consistently across records, datasets, and reference resources. |
Accuracy | Evaluate whether recorded measurements and annotations adequately represent the underlying biological observations. |
Uniqueness | Identify duplicate records, repeated identifiers, and redundant observations that could distort analysis. |
Provenance | Verify where the data originated and which transformations or processing steps were applied. |
Computational Data Lifecycle
Data acquisition from experiments, repositories, or computational predictions.
Data validation and quality assessment.
Conversion into appropriate computational representations.
Storage in files, databases, or structured data systems.
Preprocessing, normalization, and transformation where scientifically justified.
Computational analysis and statistical evaluation.
Visualization and biological interpretation.
Documentation, preservation, and responsible data sharing.