Functional genomics studies the interaction between genotype and phenotype by looking at the large numbers of variants, potentially across the whole genome (rather than at a single gene or a handful of genes). This increase in scale enables biologists to identify potential causes of disease much faster, but it has some limitations. One of the most pressing limitations is that functional genomics is confined to researching the phenotypes that one can study in a high-throughput way. Functional genomics studies typically focus either on simple phenotypes, like gain-of-function/loss-of-function and cell death, or on molecular phenotypes, like levels of protein-protein interaction and gene expression.
To study more complex and subtle phenotypes (like axon integrity) with functional genomics, researchers usually have two options. The first is to find a clever way to link the phenotype to a more simplistic one, which isn’t always possible and diminishes the ability to determine the strength of an association between genotype and phenotype, since a varied phenotype is reduced to a simple yes/no. The other option is to use a set of molecular markers (genes, proteins, enzymatic byproducts) as a proxy for the phenotype. This approach relies on a molecular understanding of the phenotype, which is usually incomplete, so there is the potential of identifying variants that affect the proxy but not the phenotype and missing variants that affect the phenotype but not the proxy. All of these problems could be solved if you could observe the phenotype directly.
FIVE (Functional Imaging for Variant Elucidation) was created to address this problem. We combine functional genomics with imaging techniques to find variants which cause complex phenotypes that are visible under a microscope.
Our Approach
We start with a list of variants of unknown significance that have some link to a specific phenotype. This list can come from exome sequencing, saturation mutagenesis, etc.
We then infect cells so that they HARBOR the variant. We do this as a pooled infection instead of an arrayed infection (1 variant per well) because pooled infections are very scalable and offer the possibility to study combinations of variants.
Once the perturbed cells are identified, they are physicially captured on a single-cell basis.
Finally, we sequence the perturbed cells to recover which variants were present (and are likely sufficient for inducing the observed phenotype).