Genetic engineering (genetic modification/manipulation) is the technology-driven alteration of an organism’s genes by adding, removing (knockout), or editing DNA. It enables gene transfer within or across species boundaries and can insert DNA randomly or at targeted genome sites, yielding organisms with novel or improved traits.

New DNA for gene editing can be obtained in two main ways. Scientists either isolate a natural gene of interest from an organism and amplify it using recombinant DNA techniques, or they order a fully designed DNA sequence from a chemical synthesis company. These fragments are then combined into a custom DNA “construct,” engineered with control elements so that, once delivered into a host cell, the new sequence is expressed and regulated in a precise, predictable way.

An organism produced via genetic engineering is genetically modified (GM), and the organism is a GMO. If DNA from another species is added it is transgenic; if DNA comes from the same species or cross-compatible species it is cisgenic; and if DNA is removed to disable a gene, the result is a knockout organism.

Unlike conventional breeding (many crosses plus selection for phenotype), genetic engineering transfers a specific gene directly, making it faster and more precise. It can introduce genes from any organism, even across domains, and reduces the chance of co-transferring unwanted genes that commonly “hitchhike” in breeding.

Key early events include Paul Berg’s first recombinant DNA molecule in 1972 (SV40 DNA combined with lambda DNA), Boyer and Cohen’s first GMO bacterium in 1973, and Rudolf Jaenisch’s first GM animal (mouse) in 1974. These advances set the stage for biotech research and commercialization.

Genentech, founded in 1976, helped launch commercial genetic engineering and produced human proteins; engineered human insulin was produced in 1978 and commercialized in 1982. GM foods reached markets in 1994 with the Flavr Savr tomato (delayed ripening/longer shelf life), while GloFish became a pet GMO in 2003.

Later milestones include the 2010 creation of a synthetic genome inserted into a bacterial cell (J. Craig Venter Institute) and the 2012 development of CRISPR/Cas9 by Doudna and Charpentier. CRISPR enables relatively easy, programmable genome changes in many organisms and accelerates targeted editing work.

In research, GMOs help study gene function/expression via loss-of-function and gain-of-function experiments, tracking, and expression studies. Knocking out disease-relevant genes can produce animal models that mimic human disorders, enabling mechanistic studies and preclinical testing of therapies in controlled systems.

Genetic engineering supports production of hormones, monoclonal antibodies, vaccines, and other biologics; mRNA vaccines are also produced using engineered nucleic acids. It also underpins gene therapy, which aims to treat genetic disease by replacing or supplying functional genes, though approaches vary in risk and ethics.

Engineered microbes or cell lines can be transformed to overexpress useful proteins (e.g., enzymes) at industrial scale using fermentation and downstream purification. Chinese hamster ovary (CHO) cells are widely used for complex biopharmaceuticals; similar methods yield enzymes used in detergents, cheese-making, and other products.

Most commercial GM crops focus on insect resistance or herbicide tolerance, though virus/fungal resistance, stress tolerance, and nutrition enhancement are also pursued. Examples include Flavr Savr tomato (shelf-life), Amflora potato (industrial starch profile), and salmon engineered with growth hormone to reach market size faster.

Creating a GMO generally involves choosing a candidate gene aligned to a desired trait, isolating or synthesizing it, assembling it into a construct, delivering it into host cells, regenerating an organism from transformed cells, and verifying insertion and expression. Screening tools (genomics/transcriptomics) aid gene discovery.

Delivery approaches include bacterial uptake (natural competence or induced by heat/electric shock), plasmid-based transformation, microinjection into animal cells, and viral vectors. Plants are commonly transformed via Agrobacterium T-DNA vectors or by biolistics (DNA-coated metal particles), with electroporation used in some contexts.

After transformation, experiments confirm the new DNA’s presence and properties using PCR, Southern hybridization, and DNA sequencing to check location and copy number. Because DNA presence doesn’t ensure function, expression is assessed by measuring RNA/protein or phenotype using methods such as qRT-PCR, Western blot, ELISA, and imaging.

Regulation emerged from the 1975 Asilomar guidelines and later frameworks including the Cartagena Protocol (adopted 2000; 157 member countries). Controversies center on labeling, gene flow, non-target impacts, patents and food-supply control. A scientific consensus holds current GM foods are not riskier than conventional, but are assessed case-by-case.

Gene editing changes DNA at specific loci. Classic editing uses programmable nucleases (e.g., Cas9) to create a targeted double‑strand break (DSB) that is repaired by cellular pathways, yielding insertions/deletions or precise sequence changes. Newer editors avoid DSBs: base editors chemically convert single bases, and prime editors write small edits using a reverse transcriptase, often reducing unwanted rearrangements.

CRISPR-Cas9 typically includes Cas9 nuclease plus a single guide RNA (sgRNA) that fuses crRNA and tracrRNA. The sgRNA contains a ~20-nt “spacer” complementary to the target DNA. Cas9 also requires a short PAM motif adjacent to the target (e.g., SpCas9 recognizes 5′-NGG-3′). PAM recognition helps Cas9 scan genomes and prevents self-targeting in bacteria.

1) Cas9–sgRNA forms a ribonucleoprotein (RNP). 2) The complex binds DNA transiently while searching for PAMs. 3) PAM binding triggers local DNA unwinding. 4) The sgRNA spacer base-pairs with the target strand, forming an R-loop (RNA–DNA hybrid plus displaced non-target DNA). Stable pairing, especially near the PAM-proximal “seed” region, licenses cleavage.

Cas9 has two nuclease domains: HNH cleaves the DNA strand complementary to the guide, and RuvC cleaves the non-complementary strand. With SpCas9, the DSB typically occurs ~3 bp upstream of the PAM. Mutating one domain creates a nickase (single-strand cut), while mutating both yields catalytically dead Cas9 (dCas9) for gene regulation and epigenome editing.

NHEJ is the dominant DSB repair pathway in many cells and can occur throughout the cell cycle. It ligates broken ends with minimal or no homology, often producing small insertions/deletions (indels). In gene editing, NHEJ is leveraged for gene knockouts by creating frameshifts or disrupting critical coding/regulatory sequences, but outcomes are heterogeneous across alleles and cells.

HDR uses a homologous DNA template to precisely repair a DSB, enabling defined “knock-in” edits when an exogenous donor is supplied. HDR is most active in S/G2 phases and competes with NHEJ, so efficiency is often lower than knockout editing. Donors can be ssODNs for small edits or plasmid/AAV templates for larger inserts, with homology arms flanking the edit site.

An alternative repair route, MMEJ (also called alt-EJ), aligns short microhomologies (often 5–25 bp) near the break before joining. This typically yields predictable deletions spanning the microhomology region and can increase large deletions or rearrangements. In CRISPR experiments, recognizing MMEJ signatures helps explain recurring deletion patterns and informs strategies to bias repair toward desired outcomes.

Effective sgRNAs balance target specificity and activity. Key considerations include PAM availability, avoiding guides with many close genomic matches, and favoring sequences with strong on-target activity (often predicted by ML scoring). The PAM-proximal seed region is especially sensitive to mismatches. For coding knockouts, guides targeting early constitutive exons can maximize frameshift likelihood across isoforms.

sgRNAs can be delivered as DNA (plasmid), RNA, or as preformed Cas9 RNP. RNP delivery often produces fast editing with shorter intracellular exposure, which can reduce off-target cutting. Synthetic sgRNAs may include 2′-O-methyl and phosphorothioate terminal modifications to improve stability and reduce innate immune activation, especially in primary cells, while maintaining on-target editing efficiency.

Off-target editing occurs when Cas9 binds and cleaves sequences similar to the intended target. Risk increases with guides that have close genomic analogs, tolerate mismatches, or target repetitive regions. Off-target outcomes include indels, larger deletions, inversions, and rare translocations between simultaneous DSBs. Chromatin context and cell type also affect binding and cutting propensities.

Common mitigations include: using high-fidelity Cas9 variants engineered to require more stringent guide–DNA pairing; shortening guide spacers in some contexts; using paired nickases to require two nearby recognition events; delivering Cas9 as RNP to limit exposure time; and careful computational screening of candidate guides. Empirical validation remains important because prediction alone may miss cell-specific off-target sites.

Specificity can be assessed by computational similarity searches, targeted deep sequencing of predicted sites, and genome-wide methods that detect cleavage footprints. Examples include GUIDE-seq (captures dsODN tags at breaks in cells), CIRCLE-seq and SITE-seq (in vitro cleavage profiling), and DISCOVER-seq (tracks DNA repair protein recruitment). Using multiple methods improves confidence in safety claims.

Base editors fuse a catalytically impaired Cas protein (often nickase Cas9) to a deaminase. Cytosine base editors can convert C→T (or G→A on the opposite strand) within an “editing window,” while adenine base editors can convert A→G. They avoid DSBs but can introduce bystander edits within the window and have distinct off-target profiles, including possible RNA editing by some deaminases.

Prime editors couple Cas9 nickase to a reverse transcriptase and use a prime editing guide RNA (pegRNA) that both targets the site and encodes the desired edit. After nicking, the pegRNA primes reverse transcription to write new DNA into the genome, enabling small insertions, deletions, and substitutions without DSBs or donor templates. Efficiency varies by locus and may require nicking the opposite strand (PE3).

In research, CRISPR is widely used to knock out genes to study function, tag proteins by HDR-mediated insertion (e.g., fluorescent tags), and perform pooled screens linking genotype to phenotype. Clinically, ex vivo editing of hematopoietic stem cells and T cells is a major focus, with aims such as reactivating fetal hemoglobin for sickle cell disease or engineering immune cells to better target cancer, emphasizing stringent specificity and safety testing.

Gene editing changes DNA at specific loci. Classic editing uses programmable nucleases (e.g., Cas9) to create a targeted double‑strand break (DSB) that is repaired by cellular pathways, yielding insertions/deletions or precise sequence changes. Newer editors avoid DSBs: base editors chemically convert single bases, and prime editors write small edits using a reverse transcriptase, often reducing unwanted rearrangements.