06. Workflows Built by Genes Together

A pathway is a flow in which multiple molecules interact in sequence to perform one cellular task. If one gene is like one function, a pathway is closer to a request-processing flow that connects several functions.

Question for this lesson: How do multiple genes connect to produce one cellular response?

Start by fixing the shape in code. The following pseudocode does not run a real cell; it only transfers the input-output relationship of insulin signalling.

def respond_to_insulin(insulin, cell_state):
    signal = insulin_receptor.detect(insulin)
    relayed = signaling_network.propagate(signal, cell_state)
    return glucose_transport.adjust(relayed)

insulin is the input and a change in glucose uptake is the output. The receptor and several signalling proteins form the intermediate processing stages. A pathway groups this input, connected processing, and cellular response into one functional unit.

A scene where insulin signaling begins

When blood glucose rises after a meal, insulin binds to a receptor on the cell surface. The receptor passes a signal to proteins inside the cell, and each activated protein activates the next one. The final response is an increase in glucose uptake by the cell.

The receptor alone cannot produce this result. The molecule that receives the input, the molecules that transmit the signal, and the molecules that carry out the response must all work together. This entire connection is called the insulin signaling pathway.

Insulin input → receptor detects it → intracellular signal transmission → increased glucose uptake

How it resembles a call flow in code

Think of a web request passing through a router, service, and database before returning a response. If one function in the middle fails, the final response changes. The system may also behave differently if one function is called far too often.

Likewise, a genetic variant can prevent one protein in a cell from working, or a particular protein can be produced in excess, changing the pathway’s final response. In cancer, this can appear as a cell-division signal that remains switched on or a process that removes damaged cells shutting down.

A gene itself is not an executable function. A gene contains information for making a protein, and the proteins produced from genes usually detect and relay the signal. The gene = function analogy borrows only the structure in which changing one component can change the full flow.

A call order in code is usually explicit, but real pathways branch and merge, share molecules with other pathways, and run many reactions concurrently. Concentration and location can also change the outcome. A pathway diagram is not a fixed execution log; it is a map of interactions known so far.

A pathway is not necessarily linear

Pathways commonly take three forms.

• Metabolic pathway: an enzyme converts molecule A into B, then another enzyme converts B into C. Molecules change in stages, as in glycolysis.
• Signaling pathway: a receptor detects a signal outside the cell and transmits it inside. Insulin signaling belongs to this category.
• Regulatory pathway: transcription factors and regulatory proteins turn multiple genes on or off. These pathways regulate programs such as cell division and cell death.

In practice, one pathway may include all three forms. It is therefore safer to check “what goes in, which molecules connect, and what the final response is” before focusing on the pathway’s name.

A pathway and a gene set are different

Experimental data does not contain the pathway itself. RNA-seq directly measures expression for each gene. Analysis tools compare those measurements with a gene set, a list of gene names known to be related to a pathway.

Suppose a pathway’s gene set contains 30 genes. If many of them are highly expressed together in a patient sample, an analysis tool may assign a score indicating that “the signal related to this pathway is strong.”

Map of biological interactions = pathway
List of related genes used for analysis = gene set

This is like the difference between a graph database and a list of search tags. A pathway can contain edges and directions, such as A activates B and B inhibits C, while a gene set retains only membership, {A, B, C}. The set cannot recover execution order, activation or inhibition, or molecular location.

Different databases can define different pathway boundaries and gene-set contents. Do not assume that pathways with the same name contain exactly the same genes.

The GSE251845 GO and GSEA exercise compares the same DEG result with GO and Reactome gene sets, then groups overlapping terms into a coherent biological signal.

What “the pathway is activated” precisely means

In RNA-seq, a high pathway score usually means that related genes are highly expressed together or are concentrated near the top of an expression ranking. It is not direct evidence that every protein in the pathway was actually active.

A pathway score is therefore not a trace from a running service. It is an indirect signal that configuration and logs related to a module were observed more often than usual. RNA-seq does not contain function-call times or return values.

Proteins may need regulation such as phosphorylation even after they are produced. RNA expression alone therefore cannot establish enzyme activity, the speed of signal transmission, or cause and effect. When needed, these are checked with proteomics, phosphoproteomics, or functional experiments.

This sequence is useful when reading analysis results:

1. What biological task does this pathway represent?
2. Which genes are in the gene set that produced the result?
3. Did a few very highly expressed genes drive the score, or did many genes move together?
4. Was actual activity confirmed with data other than RNA expression?

Pathway databases you will often encounter

• Reactome: provides human-curated reactions and signaling flows step by step.
• KEGG Pathway: provides maps of metabolic, signaling, and disease-related pathways.
• MSigDB: a collection of gene sets ready for use with GSEA and ssGSEA.
• Gene Ontology: rather than mapping pathways, it classifies biological processes, molecular functions, and cellular locations hierarchically.

Summary

A pathway is not a description of one gene’s function. It is a workflow in which multiple molecules connect to produce a cellular response. RNA-seq analysis does not observe this flow directly. It estimates it indirectly from the expression of related gene sets.

Use pathway results as a map for deciding which biological programs to investigate further, not as immediate proof of a cause or of actual activity.