How Scientists Are Learning to Rewrite DNA
CISRED explores how CRISPR, base editing, prime editing, sequencing, and computational genomics are transforming the way scientists read, modify, and understand DNA.
How Scientists Are Learning to Rewrite DNA
From reading the genome to programming its sequence, modern biotechnology is transforming DNA from something scientists observe into something they can increasingly interrogate, modify and understand.
The genome is more than a sequence.
DNA is often introduced as a long molecular code made from four letters. But modern genomics has revealed that biological information is not contained only in the letters themselves.
The same four nucleotides — adenine, thymine, cytosine and guanine — can participate in completely different biological programs depending on where they occur, how they are regulated, which proteins interact with them, and how the surrounding chromatin is organized.
A gene therefore cannot be understood simply as a line of letters. Regulatory elements, promoters, enhancers, chromatin structure and epigenetic modifications all influence whether genetic information becomes biological activity.
Before editing DNA, scientists had to learn how to read it.
The modern genome-editing revolution depends on a much older technological revolution: the ability to determine DNA sequence accurately and at scale.
Sequencing technologies transform molecules into data. Instead of looking directly at a chromosome, researchers can obtain millions or billions of sequence observations and reconstruct the genomic information computationally.
This creates the first important transition in modern biotechnology: biology becomes measurable as information.
How does CRISPR find one sequence among millions?
The central idea behind CRISPR-Cas9 is programmable molecular recognition. A guide RNA helps direct the Cas protein toward a complementary DNA sequence.
Target recognition is not simply a matter of finding a perfect twenty-base sequence. Cas9 also examines a nearby protospacer adjacent motif, or PAM. This creates a molecular search process in which genomic sequence, guide complementarity and PAM recognition work together.
Cutting DNA was only the beginning.
Early CRISPR genome editing relied heavily on creating a double-strand break and allowing the cell to repair it.
Once Cas9 generates a DNA break, the cell's own repair machinery determines much of the final outcome. Non-homologous end joining can introduce insertions or deletions, while homology-directed repair can use a donor template to introduce a desired sequence.
Target
The guide RNA brings the CRISPR-Cas complex to a selected genomic region.
Break
A nuclease can generate a double-strand DNA break at the target locus.
Repair
Cellular repair pathways determine which genetic outcome becomes established.
| Approach | Main event | Typical purpose |
|---|---|---|
| CRISPR nuclease | DNA cleavage | Gene disruption or targeted modification |
| Base editing | Base conversion | Selected nucleotide substitutions |
| Prime editing | Programmed DNA writing | Substitutions, insertions and deletions |
What if we don't want to break both DNA strands?
Base editing introduced a different idea: change individual nucleotide identities without relying on a conventional double-strand break.
Cytosine and adenine base editors combine a programmable DNA-targeting system with enzymes that chemically modify particular nucleotides. This allows defined transition changes such as C→T or A→G, depending on the editor.
The concept is important because it shifts genome engineering from cutting DNA toward directly manipulating the information encoded within individual bases.
Prime editing: when the guide RNA becomes a blueprint.
Prime editing extends the concept of programmable editing by allowing a guide RNA to encode information about the desired genetic change.
The prime editor combines a Cas9 nickase with a reverse transcriptase and a specialized prime-editing guide RNA. Instead of simply directing a cut, the guide participates in specifying the sequence that should be written.
This approach opened a broader editing space, including many substitutions as well as small insertions and deletions, without requiring a conventional double-strand break or a separate donor DNA template.
+
REVERSE
TRANSCRIPTASE
How do scientists know the edit actually worked?
Designing an editor is only half of the experiment. The other half is measuring what happened to the genome.
Researchers can sequence the targeted region to determine whether the intended modification is present. Depending on the scientific question, deeper sequencing approaches can reveal mixtures of edited and unedited molecules, unexpected substitutions, insertions or deletions, and other sequence-level outcomes.
The experiment therefore becomes a loop: design, edit, measure, analyze, redesign. Modern genome engineering increasingly depends on this feedback cycle.
Design
Choose a target and editing strategy.
Edit
Deliver and activate the molecular system.
Sequence
Generate molecular evidence from the edited genome.
Quantify
Measure editing outcomes and unwanted variants.
Interpret
Connect sequence changes to biological meaning.
The problem of the wrong target.
A genome contains billions of bases. A guide designed for one sequence may encounter other genomic regions with partial similarity.
These unintended interactions are commonly described as off-target activity. Their frequency and biological consequences depend on the guide sequence, Cas variant, genomic context, delivery conditions and other experimental factors.
Computational prediction can identify candidate sites by comparing the intended target against the genome. But prediction is not the same as measurement. Experimental validation remains essential when the consequences of unintended editing matter.
From genome editing to computational genomics.
Once genome engineering produces data, computation becomes part of the biological experiment rather than a separate activity.
Researchers can compare sequences, rank potential targets, predict off-target sites, quantify editing outcomes and model relationships between genomic variation and biological function.
More advanced computational approaches are beginning to combine sequence information with genomic context, structural information and experimental measurements. This is moving genome engineering from a sequence-only problem toward a multidimensional biological modeling problem.
The future is not simply editing DNA.
The deeper transformation is the ability to connect genomic sequence, molecular regulation, experimental measurements and computation into one continuously improving scientific system.
CRISPR made programmable genome manipulation possible. Base editing reduced the need for DNA cutting. Prime editing expanded the space of programmable changes. Sequencing made outcomes measurable. Computational genomics made those outcomes interpretable.
The next generation of biotechnology will increasingly depend on combining all of these layers — not only to change genomes, but to understand why biological systems behave the way they do.