Biological data analysis and network science · BEGINNER LESSON
Network analysis
Network analysis is a way to study a system by representing items as nodes and their relationships as edges, then examining the pattern of connections. In biology, it is often used to understand how genes, proteins, biomarkers, or disease features relate to each other and to summarize complex disease processes.
What it is and what is measured
Network analysis turns a set of biological items into a network: nodes are the items, and edges are the links between them. In this lesson, the “items” can be genes, proteins, metabolites, disease-related phenotypes, or other biomarkers. A biomarker is a measurable characteristic that can indicate a normal process, a disease process, or a response to exposure or intervention. So network analysis measures relationships among measured features, not just the features one by one. In biology, those relationships can include co-expression, physical protein interaction, or positive and negative correlation across samples.
[1] Networks as Biomarkers: Uses and Purposes - PMC [2] NetworkAnalyst - integrative approaches for protein ... - PMC - NIH
How it works
A common workflow is to choose genes or proteins of interest, build a network, then analyze and visualize it. Two main analyses are often used. Topology analysis asks which nodes are structurally important in the whole network; highly connected nodes are called hubs. Module analysis looks for small, densely connected groups called modules, which may show stronger activity. Some network tools also look at degree, the number of connections a node has, and betweenness centrality, which counts how many shortest paths pass through a node. These summaries can help reveal system-level patterns that are hard to see from individual measurements alone.
[2] NetworkAnalyst - integrative approaches for protein ... - PMC - NIH [3] Editorial: Network Bioscience - PMC - NIH [1] Networks as Biomarkers: Uses and Purposes - PMC
Why biologists use it
Biological data often come in many forms, such as DNA sequencing, blood measurements, imaging features, and electronic medical records. Network analysis can place these different measurements into one system and help explain disease complexity. It is used for gene expression data, protein–protein interaction data, disease gene prediction, and studies of relationships among different biomarkers. The goal is often hypothesis generation: the network suggests ideas for later testing rather than proving them by itself.
[1] Networks as Biomarkers: Uses and Purposes - PMC [2] NetworkAnalyst - integrative approaches for protein ... - PMC - NIH [3] Editorial: Network Bioscience - PMC - NIH
Limits and cautions
Network analysis has important limits. Many protein interaction data sets come from high-throughput experiments that can contain false positives, so results need further validation. Biological networks also change with cell type and condition, so a static network may miss interactions that only appear in a specific state. Very large networks can become hard to interpret visually, producing a “hairball” effect. For these reasons, network analysis is best treated as an exploratory tool, not final proof.
[2] NetworkAnalyst - integrative approaches for protein ... - PMC - NIH [1] Networks as Biomarkers: Uses and Purposes - PMC
Key terms
Directed network: a network where edges have a direction. Undirected network: a network where edges do not have direction. PPI network: protein–protein interaction network, a graph of proteins linked by interactions. Omics: large-scale biological data types such as genomics or proteomics.
[1] Networks as Biomarkers: Uses and Purposes - PMC [2] NetworkAnalyst - integrative approaches for protein ... - PMC - NIH [3] Editorial: Network Bioscience - PMC - NIH
SEE THE IDEA
How biological network analysis works
Genes, proteins, biomarkers, or disease phenotypes are collected from data.
Features become nodes; relationships such as interaction or correlation become edges.
Researchers look for hubs, modules, and other structural patterns.
Patterns are used to form hypotheses about mechanisms or disease states.
Data are converted into a network, then the connection pattern is studied to summarize system behavior.
A biological network helps reveal relationships among many measured features at once.
TRY IT YOURSELF
Test your understanding
Choose an answer, then check the explanation.
PUT IT TOGETHER
Match each concept to its meaning
Drag a label onto its description, or select a label and then select a description. This also works with a keyboard.
4 pairs to match
Sources and further reading
This general lesson was generated from public source material. It explains a concept; it does not interpret a private paper or provide medical advice.