Functional genomics links a gene to a function by perturbing it, then observing the resulting phenotype. Two major families of tools lend themselves to this: genome editing (CRISPR-Cas9, which modifies DNA) and RNA interference (siRNA/shRNA, which reduces messenger RNA), most often delivered by lentiviral vector. The question is simple: what happens when this gene is switched off, and what do we learn about its role?
Principle and workflow
In CRISPR-Cas9, a guide RNA (sgRNA) directs the Cas9 nuclease to a target sequence where it introduces a double-strand break. Repair by non-homologous end joining (NHEJ) generates insertions/deletions that inactivate the gene — this is the knockout; a precise knock-in, by contrast, exploits homology-directed repair (HDR) from a template. In RNA interference, a small RNA (siRNA, transient; shRNA, stable) recruits the cellular machinery to degrade the target mRNA: the gene is “switched off” (knockdown) without altering the genetic code.
Delivery is by transfection (transient effect) or by lentiviral transduction (stable integration, durable expression of Cas9/sgRNA or of the shRNA). In all cases, loss of function is validated — western blot, qPCR, sequencing — before the phenotypic analysis, a step that determines the reliability of the conclusions.
Variants and options
Beyond the knockout of a single gene, the pooled CRISPR screen introduces a library of thousands of sgRNAs into a population — each cell then carries a different knockout — then identifies, by selection and sequencing, essential genes or vulnerabilities (MAGeCK-type analysis). Other variants avoid cutting the DNA: base editing and prime editing (modifications without a double-strand break), or CRISPRi/CRISPRa (transcriptional repression or activation). On the RNA-interference side, siRNA offers transient silencing, lentiviral shRNA a stable one. Screens can be run arrayed (gene by gene) or pooled.
When and why these techniques
These approaches are used to establish causality (does this gene cause this phenotype?), validate a therapeutic target, or discover vulnerabilities at scale through screening. The choice between editing and interference depends on the objective: CRISPR delivers permanent, complete inactivation; RNA interference, a partial and tunable silencing, valuable when a total knockout would be lethal or to titrate a dose-dependent effect.
None of these tools is without trade-offs. For CRISPR, several limits must be controlled: off-target effects (cuts at undesired sites), incomplete or mosaic knockout (in-frame indels preserving function), and above all the toxicity of double-strand breaks, which trigger a p53 response in primary or non-transformed cells and can bias screens; HDR knock-in remains inefficient, and permanent inactivation is problematic for essential genes. For RNA interference, knockdown is incomplete and variable (residual mRNA remains), with sequence-related off-target effects, and transient in the case of siRNA. To circumvent the toxicity of cuts, one can turn to CRISPRi (repression without a break) or to inducible systems; lentiviral delivery, finally, imposes biosafety precautions and a semi-random integration that must be documented.
Inovarion’s expertise
Inovarion uses functional genomics to validate targets and dissect mechanisms, in oncology as in immunology. Its published work has used CRISPR-Cas9 knockout — coupled with a conditional deletion in vivo — to demonstrate that an endosomal compartment is essential to T-cell receptor signalling and to the anti-tumour response; a pooled CRISPR-Cas9 screen, from a library of chromatin regulators, to identify a vulnerability of metastatic uveal melanoma and validate it as a therapeutic target; and RNA-interference silencing to dissect the role of a breast-cancer biomarker in cell migration and the response to paclitaxel. From the perturbation strategy — editing, screening or RNA interference, and its delivery mode — through to validation of the loss of function and phenotypic analysis, each step is dictated by the biology to be interrogated.
See also: Single-cell RNA-seq (transcriptional readout of perturbations) ; In vivo & preclinical models.
Key publications
- Evnouchidou et al. IRAP-dependent endosomal T cell receptor signalling is essential for T cell responses. Nature Communications, 2020. PubMed
- Krossa et al. SETDB1 is critically required for uveal melanoma growth and represents a promising therapeutic target. Cell Death & Disease, 2025. PubMed
- Rodrigues-Ferreira et al. ATIP3 deficiency facilitates intracellular accumulation of paclitaxel to reduce cancer cell migration and lymph node metastasis in breast cancer patients. Scientific Reports, 2020. PubMed