- Metadata is one of the fundamental concepts in information science, computer science, bioinformatics, biological databases, and data management. It plays an essential role in describing, organising, identifying, searching, interpreting, and managing data.
- In biological research, scientists generate enormous amounts of information, including DNA sequences, RNA sequences, protein structures, gene expression measurements, genomic variants, experimental observations, and biological images. However, the data themselves are not always sufficient to understand their biological meaning. Additional information is needed to explain what the data represent, where they came from, how they were generated, which organism they belong to, and under what experimental conditions they were obtained. This additional information is known as metadata.
- Metadata makes biological data understandable, searchable, interpretable, reusable, and suitable for integration with other datasets. It is therefore an essential component of modern biological databases, bioinformatics workflows, scientific data repositories, and computational biology.
- Definition of Metadata
- The word metadata means data about data. It refers to information that describes, explains, identifies, or provides context for another dataset or information resource.
- In simple terms, data represent the information being studied, whereas metadata describe the characteristics and context of that information.
- For example, consider a DNA sequence: ATGCGTACGTAGCTAG
- This sequence contains biological information, but by itself it does not explain:
- Which organism does the sequence belong to?
- Which gene or genomic region does it represent?
- From which biological sample was it obtained?
- What is its sequence length?
- When was it submitted to a database?
- Who generated or submitted the sequence?
- Which sequencing technology was used?
- What is the biological function of the sequence?
- Additional information answering these questions constitutes metadata.
A simplified representation might look like this:
| Data attribute | Example |
| DNA sequence | ATGCGTACGTAGCTAG |
| Organism | Homo sapiens |
| Gene | Example gene |
| Sequence length | 16 bp |
| Sample identifier | Sample_001 |
| Sequencing method | Illumina sequencing |
| Submission date | 2026-11-07 |
| Database accession | Example accession |
- The DNA sequence is the primary biological data, while the associated descriptive information provides its metadata.
- An important distinction is that metadata are not necessarily separate from the dataset. They may be stored within the same file, in database tables, in separate metadata records, or in accompanying documentation.
- Why Is Metadata Important?
- Metadata are essential because raw data often have limited meaning without contextual information.
- For example, a gene expression value of 25 does not provide enough information for meaningful interpretation. We need to know which gene was measured, which organism was studied, which tissue was examined, what experimental conditions were used, and which measurement units or analytical methods were applied. Metadata provides this context.
- Their importance can be understood through several functions.
- Identification: Metadata helps identify individual datasets, biological samples, genes, proteins, sequences, and research experiments.
- Description: They describe the characteristics, structure, content, and properties of datasets.
- Organisation: Metadata helps organise large collections of information into searchable and manageable structures.
- Discovery: Search engines and biological databases use metadata to locate relevant datasets based on organism, gene name, tissue, experimental method, or other characteristics.
- Interpretation: Metadata provides the biological and experimental context needed to interpret observations correctly.
- Integration: They help researchers combine datasets generated by different laboratories, instruments, platforms, and databases.
- Reproducibility: Experimental metadata document how data were generated and processed, allowing other researchers to understand and reproduce scientific investigations.
- Data sharing and reuse: Metadata makes datasets more useful to other researchers by providing sufficient information about their origin, structure, methods, and meaning.
- Without appropriate metadata, even scientifically valuable datasets may become difficult to interpret, compare, or reuse.
- Main Types of Metadata
- Metadata can be classified into several categories depending on the type of information they describe and their purpose. Although different fields use slightly different classification systems, the following categories are widely used in information management and scientific data science.
3.1 Descriptive Metadata
- Descriptive metadata provides information about the identity, content, and characteristics of a dataset or information resource. Their primary purpose is to help users identify, understand, and discover data. Examples include:
- Dataset title
- Description of the dataset
- Author or researcher name
- Organism or species
- Gene or protein name
- Keywords
- Research subject
- Publication information
- Database accession number
- For example, a dataset containing DNA sequences from different animal species may have descriptive metadata identifying the species, genes, geographical origins, and biological characteristics represented in the dataset.
- Descriptive metadata is particularly important in biological databases because researchers frequently search for information using gene names, organism names, protein identifiers, and biological functions.
3.2 Structural Metadata
- Structural metadata describe how data are organised and how different components of a dataset relate to one another.
- They explain the internal organisation of information and the relationships between different data elements.
- Examples include:
- Organisation of database tables
- Relationships between records
- Arrangement of columns and rows
- File organisation
- Relationships between genes, transcripts, and proteins
- Organisation of experimental datasets
- Relationships between samples and experimental measurements
- Consider a gene expression dataset containing measurements from multiple genes and biological samples.
- Its structural metadata might describe:
- Rows representing genes.
- Columns representing samples.
- Each cell containing a gene expression measurement.
- Additional tables containing sample characteristics.
- Relationships connecting gene identifiers to gene annotation records.
- Structural metadata are essential for understanding how complex biological datasets are organised and how their components can be accessed and analysed.
3.3 Administrative Metadata
- Administrative metadata provides information required for managing, maintaining, accessing, and controlling data resources.
- They help database administrators, researchers, and institutions manage scientific information throughout its lifecycle.
- Examples include:
- Data owner
- Dataset creator
- Submission date
- Modification date
- Access permissions
- Copyright information
- Licensing conditions
- Data retention policies
- Repository information
- For example, a biological dataset deposited in a public repository may include the name of the submitting research institution, submission date, data ownership information, and conditions governing its use.
- Administrative metadata are particularly important for scientific data repositories and research data management systems.
3.4 Technical Metadata
- Technical metadata describe the technical properties of a dataset, including the file format, encoding, dimensions, software requirements, and methods used to generate or process the information.
- Examples include:
- File format
- File size
- File encoding
- Data dimensions
- Software version
- Instrument type
- Sequencing platform
- Data processing software
- Image resolution
- Compression method
- For example, a genomic sequencing dataset may include information about the sequencing instrument, read length, sequencing technology, file format, and software used during processing.
- Technical metadata helps researchers determine whether a dataset is compatible with their analytical tools and computational workflows.
3.5 Provenance Metadata
- Provenance metadata describes the origin and history of data. They explain where data came from, who generated them, what procedures were used, and how they were transformed or modified over time.
- Provenance is especially important in biological research because experimental results may pass through several stages of data generation, processing, analysis, and interpretation.
- Examples include:
- Original biological sample
- Research laboratory
- Data generator
- Experimental procedure
- Sample collection date
- Sequencing or measurement method
- Data processing steps
- Software and algorithm versions
- Previous datasets used
- Data transformation history
- For example, an RNA sequencing dataset may have provenance information describing tissue collection, RNA extraction, library preparation, sequencing, quality control, read alignment, and gene expression quantification.
- This information helps researchers understand how the final dataset was produced and assess whether the analytical results are reliable.
3.6 Semantic Metadata
- Semantic metadata describes the meaning of data and the relationships between biological concepts.
- They help computers and researchers understand what particular data elements represent, rather than simply identifying their names or positions.
- For example, the term BRCA1 identifies a particular gene, but semantic metadata can additionally describe its biological role, associated molecular functions, cellular locations, and relationships with other biological entities.
- Semantic metadata may use:
- Controlled vocabularies
- Biological ontologies
- Standardised terminology
- Gene function annotations
- Disease classifications
- Biological pathway classifications
- Molecular function descriptions
- Examples of widely used biological resources include the Gene Ontology and the Sequence Ontology.
- Semantic metadata are important for integrating information from different biological databases because different resources may use different names or terminology for related biological concepts.
3.7 Statistical Metadata
- Statistical metadata describe the statistical characteristics and analytical properties of datasets.
- They are especially important in quantitative biological research.
- Examples include:
- Mean
- Median
- Standard deviation
- Sample size
- Distribution of observations
- Missing-value frequency
- Confidence intervals
- Statistical test information
- Normalisation procedures
- Measurement units
- For example, a gene expression dataset may include metadata describing the number of biological replicates, the normalisation method, the statistical test used, and the criteria for determining differential expression.
- Statistical metadata helps researchers interpret quantitative measurements and understand the analytical procedures used to generate scientific conclusions.
Summary of metadata types
| Metadata type | Main purpose | Biological example |
| Descriptive | Identify and describe data | Gene name and organism |
| Structural | Describe data organisation | Gene expression table structure |
| Administrative | Manage and control data | Ownership and access rights |
| Technical | Describe technical properties | Sequencing platform and file format |
| Provenance | Document data origin and history | Sample collection and processing |
| Semantic | Explain biological meaning | Gene Ontology annotation |
| Statistical | Describe statistical properties | Mean expression and sample size |
- These categories are not mutually exclusive. A single biological database record may contain several types of metadata simultaneously.
- Examples of Metadata in Biological Research
- Metadata are used throughout biological research, from molecular biology experiments to genomics, transcriptomics, proteomics, and ecological studies.
- The following examples illustrate how metadata accompany different types of biological information.
4.1 Metadata Associated with DNA Sequences
- DNA sequence databases contain nucleotide sequences together with descriptive and technical information.
- A DNA sequence record may contain:
- Sequence accession number
- Organism name
- Gene name
- Sequence length
- Taxonomic classification
- Gene location
- Sequence features
- Publication references
- Submitter information
- Submission date
- For example, the GenBank database stores nucleotide sequence records with associated annotations and metadata.
- A GenBank record generally includes information about the organism, sequence definition, accession number, references, annotated features, and nucleotide sequence.
- This information allows researchers to identify sequences, investigate their biological characteristics, and retrieve relevant records.
4.2 Metadata Associated with RNA Sequencing Data
- RNA sequencing experiments generate large amounts of information about gene expression and transcript abundance.
- However, expression measurements cannot be properly interpreted without experimental metadata.
- Important RNA sequencing metadata include:
- Organism
- Tissue type
- Cell type
- Biological sample identifier
- Experimental condition
- Treatment
- Control or experimental group
- Biological replicate
- Sequencing platform
- Library preparation method
- RNA extraction method
- Read length
- Sequencing depth
- Data processing method
- For example, an RNA sequencing experiment comparing gene expression in healthy and diseased tissues requires metadata identifying the sample groups and experimental conditions.
- Without these details, researchers may be unable to determine whether observed differences are associated with disease, tissue composition, treatment, or other experimental factors.
4.3 Metadata Associated with Protein Sequences
- Protein sequence databases contain amino acid sequences and associated information describing their biological characteristics.
- Examples of protein metadata include:
- Protein name
- Protein accession number
- Organism
- Gene encoding the protein
- Protein length
- Molecular weight
- Functional annotation
- Protein domains
- Post-translational modifications
- Subcellular localisation
- Related publications
- The UniProt database provides protein sequence information together with extensive functional and biological annotations.
- Such metadata allows researchers to identify proteins, investigate their functions, compare homologous sequences, and study relationships between protein structure and biological activity.
4.4 Metadata Associated with Protein Structures
- Protein structure databases contain three-dimensional structural information about proteins and other biological macromolecules.
- Metadata associated with protein structures may include:
- Protein name
- Protein sequence
- Organism
- Experimental method
- Resolution
- Structural determination technique
- Molecular composition
- Ligand information
- Biological assembly
- Publication references
- For example, the Protein Data Bank (PDB) provides structural records containing experimental and descriptive information alongside molecular coordinates.
- This metadata helps researchers evaluate structural information and understand the conditions under which a molecular structure was determined.
4.5 Metadata Associated with Biological Samples
- Biological samples require particularly detailed metadata because their characteristics can influence experimental observations.
- Examples include:
- Sample identifier
- Organism
- Breed or strain
- Age
- Sex
- Tissue
- Cell type
- Geographical origin
- Collection date
- Storage conditions
- Treatment history
- Experimental group
- For example, a study investigating gene expression in different animal breeds should record breed identity, tissue type, age, sex, sample collection conditions, and experimental treatment where relevant.
- These details are essential for interpreting biological differences and controlling potential confounding factors.
- Metadata in Biological Databases
- Biological databases are organised collections of biological information that allow researchers to retrieve, search, analyse, and compare scientific data.
- Metadata are fundamental to their operation because they provide the information needed to organise and describe individual database records.
- Consider some major biological databases.
- GenBank: Stores nucleotide sequences with metadata such as accession numbers, organism names, sequence features, gene annotations, and references. Official GenBank website
- European Nucleotide Archive (ENA): Provides nucleotide sequence information and associated metadata about samples, studies, experiments, and sequencing runs. Official ENA website
- UniProt: Provides protein sequences and metadata describing protein identity, function, molecular characteristics, and biological annotations. Official UniProt website
- Gene Expression Omnibus (GEO): Stores functional genomics datasets with metadata describing organisms, samples, experimental conditions, platforms, and studies. Official GEO website
- Protein Data Bank: Provides three-dimensional structures and metadata describing molecular composition, experimental methods, structural resolution, and biological characteristics. RCSB PDB: Homepage Official RCSB PDB website
- These examples demonstrate that metadata are not limited to one type of biological information. They are essential across nucleotide sequence databases, protein databases, gene expression repositories, and structural biology resources.
- Metadata and Biological Data Formats
- Metadata can be stored in different ways depending on the file format and database system.
- Some formats incorporate metadata directly into the data file, while others store it in separate files or database tables.
- Understanding these differences is important when working with biological data.
FASTA format
- FASTA is a widely used format for representing nucleotide and protein sequences.
- A FASTA record typically consists of a header line beginning with > followed by the biological sequence.
| >Gene_001 Homo_sapiens chromosome_17 ATGCGTACGTAGCTAG |
- In this example:
- Gene_001 is a sequence identifier.
- Homo_sapiens identifies the organism.
- chromosome_17 describes the sequence context.
- The nucleotide string represents the biological data.
- The header contains descriptive information that functions as metadata. However, FASTA headers are not governed by one universal metadata schema, so the exact meaning and structure of header fields depend on the source.
FASTQ format
- FASTQ is commonly used to store high-throughput sequencing reads.
- A FASTQ record contains four lines:
| @Sequence_001 ATGCGTACGTAG + FFFFFFFFFFFF |
- These lines contain:
- A sequence identifier or header.
- The nucleotide sequence.
- A separator line.
- Per-base quality information encoded as characters.
- The identifier and quality information provide additional information associated with the sequence read.
- Sequencing run and sample metadata are generally maintained separately or in associated repository records.
GenBank format
- GenBank format contains nucleotide sequences together with extensive annotations and descriptive information.
- A typical record includes:
- Locus information
- Definition
- Accession number
- Version
- Organism
- References
- Feature annotations
- Nucleotide sequence
- For example, feature annotations may describe genes, coding sequences, regulatory regions, and other sequence elements.
- GenBank therefore illustrates how biological data and rich metadata can be integrated into a single record.
CSV and TSV formats
- Comma-separated values (CSV) and tab-separated values (TSV) files are commonly used to store biological measurements and associated information.
For example:
| Sample_ID | Organism | Tissue | Treatment | Expression |
| S01 | Mouse | Liver | Control | 12.5 |
| S02 | Mouse | Liver | Treated | 20.4 |
| S03 | Mouse | Kidney | Control | 9.8 |
- Here, sample identifiers, organism names, tissue types, and treatments are metadata describing the expression measurements.
- In real research datasets, it is often useful to maintain a separate sample metadata table linked to the expression matrix through unique sample identifiers.
- Metadata, Data, and Information: What Is the Difference?
- These three concepts are closely related but have different meanings.
- Data are recorded observations, measurements, sequences, or other raw or processed values.
- Metadata describes the characteristics, origin, structure, and context of data.
- Information is meaningful knowledge obtained by interpreting data in a particular context.
Consider a gene expression experiment:
| Concept | Example |
| Data | Expression value = 25.4 |
| Metadata | Gene = TP53; organism = human; tissue = liver; condition = treated |
| Information | TP53 expression was measured at 25.4 under the specified experimental condition |
- The metadata provides the context necessary to interpret the measurement.
- It is important to understand that the distinction depends on context. A sample identifier may be metadata when describing an expression dataset, but it can itself be treated as a data value when studying sample tracking or experimental design.
- Metadata Schemas and Standards
- As biological datasets are increasingly shared between laboratories, research institutions, and international databases, standardisation of metadata becomes essential.
- Different researchers may describe the same biological concept using different names, formats, abbreviations, or terminology. This creates difficulties when integrating information from multiple sources. Metadata schemas and standards help address these problems.
Metadata schemas
- A metadata schema defines the structure and organisation of metadata. It specifies which fields are required or optional, what information each field should contain, and how different elements are related.
- For example, a biological sample metadata schema might include:
| Metadata field | Description |
| Sample_ID | Unique sample identifier |
| Organism | Scientific species name |
| Tissue | Biological tissue |
| Age | Age of the organism |
| Sex | Biological sex, when relevant |
| Treatment | Experimental treatment |
| Collection_date | Sample collection date |
| Storage_condition | Conditions used for storage |
- A standardised schema ensures that researchers record similar information in a consistent manner.
Controlled vocabularies
- Controlled vocabularies are predefined sets of terms used to describe particular concepts.
- They help avoid ambiguity caused by different names for the same biological feature.
- For example, researchers may describe a tissue as “liver,” “hepatic tissue,” or “hepatic.” A controlled vocabulary can provide a standard term and identifier for the concept.
Biological ontologies
- Ontologies provide structured representations of biological concepts and the relationships between them.
- They go beyond simple lists of terms by defining how concepts are related.
- Important examples include:
- Gene Ontology (GO): Describes gene product functions, biological processes, and cellular components.
- Sequence Ontology (SO): Provides standardised terminology for biological sequence features.
- Human Phenotype Ontology (HPO): Provides standardised descriptions of human phenotypic abnormalities.
- Experimental Factor Ontology (EFO): Describes experimental factors and biomedical concepts.
- Ontologies improve the consistency and semantic interpretation of biological metadata.
Metadata exchange standards
- Several standards and initiatives support consistent metadata collection and exchange in biological research.
- Examples include:
- MIAME: Minimum Information About a Microarray Experiment, which describes essential information needed to interpret and reproduce microarray experiments.
- MINSEQE: Minimum Information about a high-throughput Nucleotide SEQuencing Experiment, which addresses reporting requirements for sequencing experiments.
- MIxS: Minimum Information about any (x) Sequence, which provides guidelines for describing sequence-based studies and their biological context.
- ISA framework: Investigation, Study, Assay, a framework for organising and reporting experimental metadata.
- Dublin Core: A general metadata standard used for describing digital resources, publications, and information objects.
- These standards improve data consistency, interoperability, and scientific reproducibility.
- Metadata and the FAIR Data Principles
- Metadata plays a central role in the FAIR principles, which provide guidelines for improving the management and reuse of scientific data.
- FAIR stands for Findable, Accessible, Interoperable, and Reusable.
- F — Findable: Metadata should make datasets easy to locate through searchable descriptions, persistent identifiers, and indexing systems.
- A — Accessible: Metadata should explain how data can be accessed, including access conditions and retrieval procedures.
- I — Interoperable: Metadata should use standardised formats, vocabularies, and identifiers so different systems can exchange and interpret information.
- R — Reusable: Metadata should provide sufficient information about data origin, methods, licensing, and quality to support future use.
- For example, a researcher depositing an RNA sequencing dataset in a public repository should provide sufficient metadata describing the organism, biological samples, experimental conditions, sequencing platform, and data processing procedures.
- This makes the dataset easier for other researchers to discover, access, interpret, and reuse.
- An important point is that data and metadata should both follow FAIR principles. In many repositories, metadata remains accessible even when the associated data are restricted.
- Metadata in Python and Biological Data Analysis
- Metadata are also essential when working with Python because biological datasets frequently contain measurements that must be connected to descriptive information.
- Python libraries such as Pandas, Biopython, and NumPy allow researchers to manage and analyse data alongside their metadata.
- Consider a small gene expression dataset:
| import pandas as pd data = { “Gene”: [“TP53”, “BRCA1”, “EGFR”], “Organism”: [“Human”, “Human”, “Human”], “Tissue”: [“Liver”, “Liver”, “Liver”], “Expression”: [12.5, 8.4, 20.1] } df = pd.DataFrame(data) print(df) |
- The DataFrame contains biological measurements and descriptive metadata.
- The columns Gene, Organism, and Tissue describe the biological context, whereas Expression contains the measured values.
- Researchers can use metadata to filter the dataset:
| liver_data = df[df[“Tissue”] == “Liver”] print(liver_data) |
- They can also calculate summary statistics:
| mean_expression = df[“Expression”].mean() print(mean_expression) |
- For more complex biological studies, metadata can be stored separately from numerical matrices and connected through sample identifiers.
- For example, a gene expression matrix may contain thousands of genes and hundreds of samples, while a separate metadata table describes each sample’s organism, tissue, treatment, age, and experimental group.
- This separation helps maintain organised and scalable data analysis workflows.
- How Is Metadata Created and Collected?
- Metadata can be generated automatically, entered manually, or obtained through a combination of different methods.
- Automatic metadata generation: Software applications, sequencing instruments, microscopes, and laboratory information systems can automatically record technical information such as timestamps, instrument settings, file sizes, and software versions.
- Manual metadata entry: Researchers may enter information such as organism, tissue type, experimental conditions, sample identifiers, and collection details.
- Database-generated metadata: Biological repositories assign accession numbers, record submission dates, maintain version histories, and generate other administrative information.
- Computationally derived metadata: Bioinformatics tools can generate metadata through computational analysis, such as sequence length, GC content, predicted gene locations, protein domains, and statistical summaries.
- Metadata extraction: Information can also be extracted from existing files, publications, database records, and laboratory documentation.
- For reliable biological research, metadata collection should ideally begin when samples are collected and continue throughout the experimental and analytical workflow.
- Metadata Quality and Common Problems
- The usefulness of metadata depends on their quality. Incomplete, inaccurate, inconsistent, or poorly structured metadata can reduce the scientific value of biological datasets.
- Several common problems occur in biological data management.
- Incomplete metadata: Important information, such as tissue type, organism, sample origin, or experimental conditions, may be missing.
- Inconsistent terminology: Different researchers may use different names for the same biological feature.
- Incorrect identifiers: Errors in gene identifiers, sample identifiers, or database accession numbers may cause incorrect data matching.
- Inconsistent formats: Dates, measurement units, organism names, and other fields may be recorded using different formats.
- Outdated metadata: Information may become inaccurate when datasets are updated or re-annotated.
- Duplicate metadata: The same sample or biological record may be represented more than once with different identifiers.
- Insufficient provenance: The absence of information about data generation and processing can make scientific results difficult to reproduce.
- Best practices for maintaining metadata quality
- Use standardised metadata schemas.
- Use accepted scientific names for organisms.
- Apply controlled vocabularies and ontology identifiers where appropriate.
- Assign unique identifiers to samples and datasets.
- Record measurement units explicitly.
- Document experimental conditions and data processing procedures.
- Distinguish missing values from zero values.
- Maintain clear documentation and version histories.
- Validate metadata before submitting datasets to public repositories.
- Keep metadata associated with the corresponding biological data throughout the research lifecycle.
- These practices improve data reliability and facilitate integration across biological databases.
- Metadata Versus Annotation: Are They the Same?
- Metadata and annotation are closely related concepts, but they are not identical.
- Metadata describe a dataset or information resource, whereas annotation involves adding explanatory information or biological interpretations to data.
- For example, consider a DNA sequence.
- Its metadata may include:
- Accession number
- Organism
- Sequence length
- Submission date
- Sequencing method
- Its biological annotations may include:
- Gene locations
- Coding sequences
- Exon boundaries
- Promoter regions
- Predicted functional elements
- Annotations describe specific features or biological meanings associated with the sequence.
- In practice, annotations can form part of a dataset’s metadata, particularly when they describe its content and biological characteristics. The distinction depends on the context and database design.
- Metadata Versus Data Documentation
- Metadata and data documentation are also related but serve different purposes.
- Metadata consist of structured information describing a dataset, whereas documentation provides more detailed explanations of how the dataset was generated, organised, processed, and analysed.
- For example, a gene expression dataset may contain metadata describing the organism, tissue, treatment, and sequencing platform.
- Its documentation may explain:
- How samples were collected.
- How RNA was extracted.
- How sequencing libraries were prepared.
- Which quality-control procedures were applied.
- How reads were aligned.
- How gene expression was quantified.
- Which statistical methods were used.
- Metadata provide structured descriptions that can be searched and processed by computer systems, while documentation provides additional explanations for researchers.
- Both are important for understanding and reproducing biological analyses.
- The Role of Metadata in Modern Biological Research
- The importance of metadata is increasing as biological research becomes increasingly data-intensive and interconnected.
- Modern research frequently involves combining information from multiple sources, including genome sequencing projects, transcriptomic experiments, protein databases, clinical studies, and molecular structure repositories.
- Metadata makes it possible to connect these different information sources and understand their relationships.
- Some particularly important applications include:
- Genomics and comparative genomics: Metadata help identify organisms, chromosomes, genomic regions, sequencing platforms, and experimental conditions, supporting comparisons between genomes.
- Transcriptomics: Metadata allow researchers to compare gene expression across tissues, developmental stages, diseases, treatments, and experimental groups.
- Proteomics: Protein metadata support protein identification, functional annotation, molecular characterisation, and comparison between organisms.
- Population genetics: Metadata describing populations, geographical locations, breeds, individuals, and sampling conditions help researchers investigate genetic diversity, population structure, and evolutionary relationships.
- Animal breeding and genetics: Metadata describing animal breeds, pedigrees, sex, age, production traits, geographical origin, and genotypes are essential for studying inherited variation and breeding performance.
- Systems biology: Metadata helps integrate genomic, transcriptomic, proteomic, metabolomic, and other biological datasets to investigate complex biological systems.
- Artificial intelligence and machine learning: Metadata provides important contextual information for selecting, organising, labelling, and interpreting datasets used to train computational models.
- Summary: Understanding Metadata Through a Simple Example
- Consider a research project investigating gene expression differences between two animal breeds.
- The researcher collects tissue samples from both breeds and performs RNA sequencing.
- The resulting dataset contains gene expression measurements, but the measurements alone are insufficient to understand the experiment.
- The researcher also needs metadata describing the biological and experimental characteristics of the samples.
Example: Gene expression study in two animal breeds
Primary biological data: 12,450
Example gene expression measurement
Associated metadata
| Metadata field | Example value |
| Sample ID | B001 |
| Organism | Bos taurus |
| Breed | Breed A |
| Tissue | Skeletal muscle |
| Age | 18 months |
| Sex | Female |
| Experimental group | Control |
| Technique | RNA sequencing |
| Sequencing platform | Illumina |
| Data format | Gene expression matrix |
What the metadata tell us
- The expression measurement belongs to a particular sample from a specified animal breed, tissue, age group, and experimental condition. The sequencing information also provides technical context for understanding how the measurement was generated.
- This example illustrates the central role of metadata: they transform isolated measurements into interpretable scientific datasets.
Conclusion
- Metadata are structured descriptive information that provide context, meaning, organisation, and traceability to data. They are essential for identifying, searching, interpreting, managing, integrating, and reusing scientific information.
- In biological research, metadata describes important characteristics of DNA and RNA sequences, proteins, molecular structures, biological samples, experimental conditions, sequencing technologies, and computational analyses.
- Different categories of metadata, including descriptive, structural, administrative, technical, provenance, semantic, and statistical metadata, serve complementary purposes. Their standardisation through metadata schemas, controlled vocabularies, and biological ontologies improves data interoperability and scientific reproducibility.
- In bioinformatics and biological databases, metadata are as important as the primary biological data because they provide the information needed to understand what a dataset represents, how it was generated, and how it can be used. As biological research increasingly relies on large-scale data integration, artificial intelligence, and computational analysis, high-quality metadata will remain fundamental to effective scientific data management.