Oncogene Dependency and Synthetic Lethality: Building the Right Model

"Dependency" and "synthetic lethality" are often used interchangeably in informal discussion, and they are not the same claim.

Dependency means the cell requires the gene in its current context. Synthetic lethality means the cell requires it because of a specific second lesion — the relationship is conditional by definition.

The distinction determines the model: dependency work needs a context-matched panel, while synthetic lethality needs a pair of contexts that differ only in the second lesion.

The third row is the one most often skipped: a knockout line that has lost the upstream pathway cannot report on a downstream dependency.

An isogenic pair differs only in the lesion of interest, which removes the confound that dominates cell-line panels: co-occurring mutations.

Design rules.

• • Both directions. Introduce the lesion into a wild-type background and remove it from a mutant background. Concordance between directions is far stronger than either alone.
• • Multiple clones per genotype. Clonal drift is real; a single pair is a single comparison.
• • Confirm the functional state, not just the genotype. A knock-in that does not activate the pathway does not create the context you intended.
• • Check for secondary adaptations. Cells engineered to carry an oncogene may acquire compensating changes during selection.

What isogenic panels cannot do. They do not model tumour heterogeneity, stroma, or immune pressure. A dependency confirmed in an isogenic panel still needs a lineage-matched tumour model before a therapeutic claim.

A synthetic lethal relationship requires three arms:

Without arms 1 and 2, an effect in the double mutant cannot be attributed to the interaction — it may simply reflect either lesion alone.

Practical designs.

• • Isogenic double mutants, built by sequential editing in one parental background.
• • Pharmacological mimicry of one lesion (for example, an inhibitor of B) in an isogenic pair differing at A — faster, but compound-specific effects enter the comparison.
• • Pooled combinatorial screens for discovery, with the same three arms reconstructed individually for validation. The mechanics of building such a screen are covered in our separate guide to synthetic lethality screens.

The first row is the standard; the second is the most sensitive; the fourth explains why the effect occurs. A dependency paper typically needs the first plus one of the others.

Deriving resistant clones under continued selective pressure remains the most convincing evidence that the effect is on-target. Sequencing the resistant clones identifies the bypass route — a second-site mutation in the target, pathway reactivation, or a parallel pathway.

Two practical notes: resistance can arise through multiple mechanisms, so a handful of clones is not a complete survey; and the absence of resistance is not proof of specificity, only that the experiment did not find it.

EDITGENE builds isogenic oncogene panels in both edit directions, lineage-matched tumour lines with documented pathway status, derived resistance clones, and matched non-targeting controls — with growth and pathway-state characterization before any dependency experiment.

Tell us the lesion and the lineage. Contact EDITGENE to request a project consultation or quote.

Dependency is context-dependent: lineage, co-occurring lesions, and culture conditions all influence it. Statements should specify the context in which the dependency was observed.

No cell model reproduces a tumour. A dependency confirmed in vitro requires lineage-matched corroboration, and synthetic lethality requires all three arms.

• • Ben-David U, et al. Genetic and transcriptional evolution alters cancer cell line drug response. Nature. 2018;560:325–330.
• • Hart T, et al. Evaluation and design of genome-wide CRISPR/SpCas9 knockout screens. G3. 2017;7:2719–2727.
• • Tsherniak A, et al. Defining a cancer dependency map. Cell. 2017;170:564–576.
• • O'Neil NJ, et al. Synthetic lethality and cancer. Nature Reviews Genetics. 2017;18:613–623.