Reference
Genome engineering
Editors that trade the double-strand break for their own constraints, minimal genomes full of unknown genes, and biosecurity's weakest assumption.
Twelve subjects that descend from one sentence: the nuclease only makes a break — what the edit becomes is decided by the cell’s repair machinery. The newer editors are variations on escaping that dependence, each buying precision at its own price: base editing works on a displaced strand and edits a wider window than you asked for; prime editing carries its own template and puts the whole engineering problem into guide design; epigenome editing changes whether a gene is read without touching what it says, and must answer whether the change lasts.
The second idea: writing genomes runs into what nobody yet understands. A bacterium reduced to 473 genes still contains dozens whose function is unknown — and that is the result, not a footnote. A circuit that worked on its own stops working inside a cell, and the synthetic-biology page counts the reasons.
The third: containment and biosecurity are rates and procedures, not features. A genetic kill switch is a rate set by how many mutants a population can afford; a gene drive breeds its own resistance allele each time it fails; DNA-synthesis screening works exactly as well as sequence similarity does.
Start with genome editing — every other page here modifies its central sentence.
- Xenobiology & expanded genetic alphabet Unnatural base pairs, xeno-nucleic acids, orthogonal translation and recoded genomes: what makes each chemically orthogonal, and why a genetic firewall is only as good as the assay that measured it.
- Base editing How a deaminase fused to a Cas nickase converts one base into another inside the R-loop, and the three limits that follow: the editing window, guide-independent deamination, and the restricted set of available conversions.
- Biosafety and laboratory biosecurity Directional airflow and pressure cascades, why the cabinet is the primary barrier and the room the secondary one, how HEPA filtration actually works, and where human factors set the real limit.
- CRISPR screening as a service Why pooled library screens work as counting experiments: cells per guide, multiplicity of infection and one guide per cell, dropout versus enrichment, and what a reporter sort measures that a viability screen cannot.
- DNA synthesis biosecurity screening Homology-based screening of synthesis orders, the 200-nucleotide threshold in US guidance, and the defeats: split orders, functional-equivalence redesign, and benchtop synthesisers outside the screened supply chain.
- Ecological engineering and gene drives How homing drives convert a heterozygote into a homozygote, why non-homologous end joining at the cut site is the dominant failure mode, and what daisy-chain and threshold designs concede about irreversibility.
- Epigenome editing dCas9 fused to a methyltransferase, demethylase or histone-modifying domain, and the heritability problem: which marks survive after the effector is gone, and which relax as soon as it is.
- Genome editing Why a programmable nuclease is a cutting tool and not a writing tool — repair-pathway choice, off-target measurement, and the on-target damage that short amplicon assays cannot see.
- Intelligent genetic switches and kill switches Why every single-mechanism biocontainment circuit is defeated by loss-of-function mutation, how synthetic auxotrophy and layered designs answer it, and how to read a reported escape frequency.
- Minimal genomes and synthetic cells Top-down genome reduction and bottom-up cell assembly: what JCVI-syn3.0 established about essentiality, why quasi-essential genes make minimality context-dependent, and what a synthetic cell still cannot do.
- Prime editing How a nickase-reverse transcriptase fusion writes a specified sequence without a double-strand break or a donor, and why primer binding site length, flap equilibrium and mismatch repair decide whether it works.
- Synthetic biology Abstraction, parts and the engineering premise of synthetic biology, set against what actually breaks it: context dependence, host burden, expression noise, and evolution acting on the construct.