Genes carry the instructions, and proteins do the work. In the simplest terms, a gene is a stretch of DNA whose sequence specifies the order of amino acids in a protein, and the cell builds that protein through a two-step relay: first copying the DNA into a messenger molecule (RNA), then reading that messenger to assemble the protein. That core relay, often called the “central dogma” of molecular biology, is real and foundational, but the full relationship between genes and proteins is far messier and more interesting than a one-gene-makes-one-protein diagram suggests.
The First Step Is Copying DNA Into RNA
Every protein-coding gene sits in a long strand of DNA inside the cell’s nucleus. When the cell needs a particular protein, it does not ship the DNA itself out to the protein-building machinery. Instead, an enzyme called RNA polymerase latches onto the gene and builds a complementary strand of messenger RNA (mRNA), reading the DNA letter by letter. This copying step is called transcription.
RNA polymerase is not a simple copier. Structural studies of the enzyme show it forms a clamp around the DNA near its active site, and once RNA is being produced, that clamp locks into a closed position, making the copying complex remarkably stable while it works its way along the gene.1PubMed. Architecture of RNA polymerase II and implications for the transcription mechanism That stability matters because human genes can be extremely long, and the enzyme needs to finish the job without falling off partway through.
The raw mRNA transcript is not ready for use right away. It undergoes processing: a chemical cap is added to one end, a tail of repeating adenine bases is added to the other, and stretches of non-coding sequence (called introns) are snipped out while the coding stretches (exons) are stitched together. Only after this editing does the mature mRNA leave the nucleus for the cytoplasm, where the protein-building machinery waits.
One Gene, Many Proteins
Here is where the textbook version gets too simple. During that splicing step, the cell does not always stitch the exons together in the same order or include the same set. By choosing different combinations of exons, a single gene can produce multiple distinct mRNA versions, each encoding a slightly different protein. This process, called alternative splicing, is the main source of protein diversity in complex organisms. Up to 95% of human genes with multiple exons undergo alternative splicing, which means the roughly 20,000 protein-coding genes in the human genome actually generate a far larger catalog of proteins.2PubMed Central. Alternative splicing: Human disease and quantitative analysis from high-throughput sequencing
Some examples of alternative splicing are staggering in scale. In fruit flies, a single gene involved in nervous system wiring can theoretically produce over 38,000 different protein variants through a form of splicing in which only one exon from a large cluster is included at a time.3PubMed. Mutually exclusive alternative splicing of pre-mRNAs That is more distinct proteins from one gene than most organisms have genes in total. The human version of the same principle is less extreme but still pervasive, and it means that “one gene equals one protein” is more of a historical simplification than a biological rule.
Building the Protein From the Message
Once a mature mRNA reaches the cytoplasm, it meets the ribosome, a large molecular machine that reads the mRNA’s sequence three letters at a time. Each three-letter group (a codon) specifies one amino acid. Small adapter molecules called transfer RNAs (tRNAs) carry the correct amino acid to the ribosome, matching their own three-letter anticodon to the mRNA’s codon. The ribosome links each incoming amino acid to the growing chain, and the protein elongates one residue at a time.
This process, translation, is tightly choreographed. Single-molecule studies have tracked individual tRNA molecules moving through the ribosome in real time, revealing that each tRNA goes through multiple intermediate positions as it delivers its amino acid, and the two tRNAs sitting on the ribosome at any given moment fluctuate between distinct physical arrangements before the ribosome steps forward to the next codon.4PubMed Central. tRNA dynamics on the ribosome during translation These structural gymnastics help ensure accuracy. A misread codon means a wrong amino acid, and even a single wrong amino acid can ruin the finished protein.
Shape Is Everything
A freshly built chain of amino acids is not yet a functional protein. It has to fold into a precise three-dimensional shape, and that shape determines what the protein actually does. An enzyme’s ability to speed up a chemical reaction, a receptor’s ability to grab a hormone, an antibody’s ability to recognize a pathogen: all depend on the protein adopting the right fold.
Folding is not left to chance. Cells maintain an elaborate quality-control network of helper proteins (molecular chaperones), folding enzymes, and degradation systems that together ensure polypeptide chains reach their correct shapes and clear away the ones that fail.5PubMed Central. Navigating the landscape of protein folding and proteostasis: from molecular chaperones to therapeutic innovations When this quality-control network breaks down, misfolded proteins accumulate, and that is a hallmark of diseases like Alzheimer’s, Parkinson’s, and ALS. So the gene-to-protein relationship does not stop at building the chain; the cell also has to supervise the folding step, and failure at that stage can be just as devastating as a mutation in the gene itself.
What Happens After a Protein Is Made
Even after a protein folds correctly, the cell keeps modifying it. Hundreds of different chemical tags can be attached to a finished protein, changing its activity, location, lifespan, or interactions with other molecules. One of the most versatile of these tags is ubiquitin, a small protein that can be attached to other proteins in a staggering number of combinations. Considering the different ways ubiquitin molecules can link to each other, plus additional modifications like phosphorylation and acetylation on ubiquitin itself, the number of possible signals is essentially unlimited.6Nature. Protein ubiquitination is a dynamic multifaceted post-translational modification involved in nearly all aspects of eukaryotic biology Ubiquitin tagging plays a role in almost every aspect of cell biology, from controlling when proteins are destroyed to activating immune signaling.
Protein destruction is itself a regulated process. The ubiquitin-proteasome system tags proteins with ubiquitin chains and then feeds them into a barrel-shaped molecular shredder (the proteasome), which breaks them back down into amino acid fragments.7Current Biology. Synaptic Plasticity: Importance of Proteasome-Mediated Protein Turnover This controlled destruction is not waste. It is how the cell rapidly adjusts protein levels in response to changing conditions, clears damaged proteins, and even manages long-term memory formation at nerve synapses.
These post-translational modifications mean that knowing a gene’s sequence tells you what protein it encodes, but it does not tell you which version of that protein is active in a particular cell at a particular moment. The same protein can behave very differently depending on which chemical tags it carries.
Why mRNA Levels Do Not Reliably Predict Protein Levels
A reasonable assumption would be that if a cell makes a lot of a gene’s mRNA, it will also have a lot of the corresponding protein. In practice, the correlation between mRNA abundance and protein abundance is notoriously poor.8PubMed. Correlation of mRNA and protein in complex biological samples This disconnect has practical consequences for researchers and clinicians who use gene-expression data to infer what is happening at the protein level.
Several layers of regulation sit between the mRNA and the final protein. Translation rates differ from one mRNA to another. Some mRNAs are translated many times before being broken down; others are translated just a few times. Proteins themselves vary enormously in how long they last, from minutes to months. And the post-translational modifications discussed above can mark a protein for rapid destruction or stabilize it for long-term use. All of these factors mean that measuring a gene’s mRNA output is, at best, a rough proxy for how much active protein the cell contains.
When a Mutation Changes the Protein
Because genes spell out proteins letter by letter, a change in even a single DNA letter can alter the protein. A point mutation that swaps one nucleotide for another may result in a different amino acid at one position in the protein (a missense mutation), or it may happen to land on a codon where the change does not alter the amino acid at all, because multiple codons can specify the same amino acid.9PubMed Central. Analyzing effects of naturally occurring missense mutations Some mutations introduce a premature stop signal, cutting the protein short. Others delete or insert bases, throwing off the reading frame so that every downstream amino acid is wrong.
The consequences depend on where the change lands. A missense mutation in a critical part of an enzyme’s active site might destroy its function entirely, while the same kind of mutation in a flexible loop on the protein’s surface might have no measurable effect. This is why genetic testing can identify a mutation but often cannot predict its severity without additional evidence about the protein’s structure and function. Sickle cell disease, for instance, results from a single amino acid swap in hemoglobin, one change out of 147 positions in that protein chain, yet it reshapes the entire red blood cell.
Genes That Do Not Make Proteins At All
For decades, the gene-protein relationship was treated as the whole story: genes exist to encode proteins. But it turns out that only about 1.5% of the human genome directly codes for protein. Much of the rest was once dismissed as “junk DNA,” yet a growing body of work shows that large portions of the non-coding genome are transcribed into RNA molecules that never get translated into protein but still have important jobs.
Non-coding RNAs come in many sizes and flavors. Small non-coding RNAs (roughly 20 to 30 nucleotides long) and long non-coding RNAs (over 200 nucleotides) have been established as key regulators of gene expression across a wide range of biological processes, from embryonic development to immune defense.10PubMed Central. Gene regulation by non-coding RNAs Long non-coding RNAs, for example, can regulate gene activity by altering how DNA is packaged, by influencing the stability of mRNAs, or by interfering with signaling pathways. Disruption of these regulatory RNAs has been linked to neurological disorders, immune dysfunction, and cancer.11PubMed Central. Long non-coding RNAs: from decades of discovery to mechanisms and functions
This means the relationship between genes and proteins is not always direct. Some genes influence proteins not by encoding them but by producing RNA regulators that dial protein production up or down, target certain mRNAs for destruction, or reshape the chromatin landscape so that other genes become accessible or silenced.
Epigenetic Controls on Gene Access
Even a perfectly intact gene cannot make a protein if the cell’s machinery cannot physically reach it. DNA in the nucleus is wound tightly around spool-like histone proteins, forming a structure called chromatin. Chemical modifications on those histones act as switches that either open the chromatin (making the gene accessible for transcription) or tighten it (silencing the gene). In all organisms with a nucleus, histone modifications play a major role in regulating which genes are active at any given time.12PubMed Central. Histone modifications and dynamic regulation of genome accessibility in plants
Epigenetic regulation explains how cells with identical DNA can produce wildly different sets of proteins. A liver cell and a neuron carry the same genome, but their histone modification patterns differ, so different genes are exposed in each cell type. Environmental signals like diet, stress, and chemical exposure can shift these patterns, which is one reason that the gene-to-protein pipeline is sensitive to context. Your genes set the menu; epigenetic marks determine which dishes get served in a particular cell at a particular time.
Prions and Protein-Only Inheritance
The standard flow of information runs from gene to RNA to protein. Prions are a disturbing exception. In prion diseases like scrapie in sheep or Creutzfeldt-Jakob disease in humans, a normal cellular protein (PrP) misfolds into an abnormal shape, and that misfolded version can force correctly folded copies of the same protein to adopt the pathological shape. The disease spreads not because a gene has mutated, but because a protein’s three-dimensional conformation propagates itself.13PubMed. Prion disease–the propagation of infectious protein topologies
Different prion strains produce distinct misfolded structures, and experiments have shown that these structural differences can self-propagate in cell-free systems, meaning the information for a prion “strain” is carried entirely in protein shape rather than in DNA sequence.14PubMed. Non-genetic propagation of strain-specific properties of scrapie prion protein Prions do not invalidate the gene-protein relationship, since the gene still encodes the protein’s amino acid sequence. But they show that information can flow at the protein level alone, with shape acting as a form of heritable instruction that bypasses DNA entirely.
How the Gene-Protein Link Is Studied Today
Historically, researchers identified proteins first and then worked backward to find the gene, or studied a gene and predicted the protein it might produce. Modern proteogenomics works both directions at once, combining genome sequencing with mass spectrometry to map proteins directly back to the genomic sequences that encode them.15PubMed Central. Proteogenomics: Integrating Next-Generation Sequencing and Mass Spectrometry to Characterize Human Proteomic Variation This approach has become increasingly powerful as sequencing costs have dropped and mass spectrometry instruments have become more sensitive.
Proteogenomic mapping can identify protein-coding regions of the genome that were not predicted by gene-finding algorithms, catching genes that computational methods missed.16PubMed Central. Whole human genome proteogenomic mapping for ENCODE cell line data: identifying protein-coding regions It also reveals which splice variants and protein modifications actually exist in a given tissue or disease state, rather than relying on predictions from the DNA sequence alone. For cancer research in particular, proteogenomics has become a critical tool for understanding how a tumor’s specific mutations translate into the abnormal proteins that drive its growth.
mRNA Therapeutics and Hacking the Relay
Understanding the gene-to-protein relay has opened the door to a striking medical strategy: skip the gene entirely and deliver synthetic mRNA directly into cells, letting the cell’s own ribosomes build the protein you want. This is the principle behind the Moderna and Pfizer-BioNTech COVID-19 vaccines, which deliver mRNA encoding the SARS-CoV-2 spike protein packaged inside lipid nanoparticles. Once inside cells, the mRNA is translated into spike protein, which triggers an immune response without any virus being present.17PubMed Central. Lipid Nanoparticle-mRNA Formulations for Therapeutic Applications
The same platform is now being explored for protein replacement therapies (delivering mRNA for a protein the patient’s own genes fail to produce), genome editing (delivering mRNA for editing enzymes), and cancer immunotherapy (delivering mRNA that teaches immune cells to recognize tumor-specific proteins). The approach works because the gene-to-protein relay is universal: give any human cell a properly designed mRNA, and it will translate that message into protein, regardless of what its own genome says. That universality, one of the most fundamental features of the gene-protein relationship, turns out to be one of the most medically useful.
Where the Gene-Protein Relationship Likely Began
The link between nucleic acid sequences and protein assembly is so central to life that its origin is one of the biggest open questions in biology. One influential model proposes that the translation system arose in stages within an ancient “RNA world,” a hypothetical era in which RNA molecules served as both information carriers and catalysts. Under this scenario, amino acids and short peptides first became useful because they enhanced the chemical reactions catalyzed by RNA enzymes (ribozymes), and the system for reading a genetic code and building proteins evolved incrementally from that starting point.18PubMed Central. On the origin of the translation system and the genetic code in the RNA world by means of natural selection, exaptation, and subfunctionalization
The ribosome itself provides a clue. Its catalytic core, the part that actually forms the chemical bond between amino acids, is made of RNA, not protein. That strongly suggests the ribosome predates the proteins it now builds, consistent with the idea that RNA-based life invented protein synthesis as a useful add-on rather than the other way around. However the details ultimately shake out, the gene-protein relationship appears to be not just a feature of modern cells but one of the oldest molecular partnerships on Earth.