Other types of data

[jhrcook]

It would be interesting to experiment with additional input data in the models. This discussion serves as a place to record different options and discuss their feasibility.

[jhrcook]

Include experimental evidence for off-target effects of guides

This may be another factor to explain some variance in the effects of sgRNAs. Bing mentioned that a group had experimentally determined off-target effects of sgRNAs. Maybe ask him what the paper was or do some research of my own. This could also be estimated by including a covariate for the number of loci that are targeted by a guide.

[jhrcook]

Incorporate mutational frequency to modeling of mutation varying effects

We can gather more information of the mutations of genes such as their mutational frequency in cancer or whether they are known oncogenes or tumor suppressors (COSMIC CGC). There should be a way of building that information into the model to help accurately measure the effect of a mutation on the LFC - perhaps an interaction covariate (is_mutated * icgc_mut_frequancy)?