Predicted CRISPR KO gene score (precursor to Predictability score)

Q&A
1

Hello. I am thinking that to generate the Predictability score the Machine Learning (ML) models first generate the gene knockout scores so that pearson correlation can be calculated. Is this correct?

If so, I am interested in looking at the gene knockout scores from the ML models. Are these available? Could you point me to where I can download it?

Thank you,

Hoang

2

Hello,

It sounds like you’re looking for the predictions generated by the random forest models whose results are summarized on the “Predictability” tab, correct?

If so, unfortunately, in the current pipeline we do not retain the predictions, only the summary scores.

We are exploring updating the pipeline which fits the random-forest models, and one change we’ve commited to in doing that rework, is saving the predictions and making them available for download and viewable in Data Explorer. However, I don’t have a concrete timeline for when the new pipeline will be completed and we’d be rolling out that functionality.

Thanks,

Phil

3

Hi Phil. That’s what I’m looking for. Thank you for letting me know. Thank you DepMap for working on making the data more accessible. I’m looking forward to it.

All the best,

Hoang