Identifying the genetic underpinnings of Toxoplasma-host interactions using advanced spatial phenotype screening

The evolutionary arms race between host and microbe has provided mammalian cells with an arsenal of molecular mechanisms to counter and clear invading microbes. However, some pathogens can evade clearance and establish lifelong chronic infection. The parasite Toxoplasma gondii infects all warm-blooded vertebrates including one third of the global human population.

While T. gondii infection is generally asymptomatic in healthy individuals, immunocompromised patients such as HIV/AIDS patients can develop toxoplasmic encephalitis.

Recently, pooled CRISPR/Cas9 loss-of-function fitness screens have accelerated the discovery of T. gondii virulence genes. However, fitness-based readouts fail to detect complex phenotypes at the subcellular level.

Advances in artificial intelligence, in particular deep convolutional neural networks (CNNs), facilitate screens of spatial phenotypes captured in imaging data.

Here, we aim to identify novel virulence genes by screening a mutant pool of CRISPR/Cas9 knockout T. gondii parasites with high content imaging and CNNs.

This screening method will for the first time allow screening hundreds of T. gondii genes for their role in spatial phenotypes.

Specifically, we will identify T. gondii genes that facilitate host microtubules and MTOC recruitment and host Golgi fragmentation.

The proposed project will grant agnostic new insight into host-parasite interactions in T. gondii infection and establish a platform to study other host-parasite interactions.