Table 1 Comparison of transposon mutagenesis and CRISPR screens in vivo

Cancer model quality

Transplantation models: limited by lack of immune system.

Autochthonous models: high quality.

High-quality given autochthonous nature.

Tissue delivery of components

Can be technically challenging. Options include:

- Transfection, with or without PB/SB: allows multiplexing; limited to certain organs.

- Lentivirus or AAV: wide range of tissues are accessible; relatively limited cargo capacity.

- Nanoparticles: high cargo capacity, low immunogenicity.

Constitutive and conditional transposase mouse knock-in alleles are available. No further delivery of exogenous components is required.

Efficiency of mutations

High efficiency in vitro (for transplantation model).

In vivo mutagenesis has typically lower efficiency than in vitro.

High efficiency of gene disruption in vivo; gene activation efficiency is variable. Therefore, TSGs are more identifiable than oncogenes¹.

Types of mutations

Disruptive (indels)/knockout.

Transcriptional activation/repression.

Translocations/deletions.

Point mutations.

Disruption or activation of expression.

Drop-out screens are not possible.

Alleles targeted

Can induce mutations in both alleles of diploid cells

Usually only one allele is mutated in diploid cells.

Unwanted genetic effects

Low off-target effects.

On-target unintended effects may occur.

Local hopping

Footprint mutations with SB.

Genome coverage in screen

Focused or genome-wide libraries.

High library coverage can be challenging to achieve in vivo.

Whole-genome including non-coding regions.

PB favors open chromatin.

SB and PB have differing integration preferences.

Time to conduct screen

Relatively short: direct injection of CRISPR components into tissues is required.

Relatively long: compound mutant mice (typically three or four alleles) need generating.

Costs of screen

Relatively low.

High costs due to long mouse breeding times.