DESKGEN Predict Library

A custom scoring function optimized for your model.
Rational CRISPR library design optimized to your model

DESKGEN Predict provides your team with novel CRISPR activity or specificity scores optimized to your laboratory workflow. Scores customized for your team outperform published algorithms by accounting for variables unique to your experimental process.


Train once - accelerate forever

To unlock DESKGEN Predict libraries, we must first understand how CRISPR behaves in your workflow by designing an initial training library, or by using your pre-existing data.

Once we have what we need, DESKGEN AI will create a new scoring function using machine learning that can be used to design additional DESKGEN Libraries for you.


Training options

DESKGEN Predict projects can take into consideration all parameters that are pertinent to your work. Whether it's a unique form of CRISPR delivery or a particular organism of interest, our team will determine which variables contribute to succesful genome editing outcomes.

Customize your Predict Library
Parameter
Customizable?
Example
Genome Yes HEK293 cell line
Experimental intent Yes Gene knockout
Nuclease Yes Cpf1
Delivery method Yes Lentivirus
Scoring algorithm Yes Activity
Data source Yes Client screening data

A Case Study

How industry researchers used a Predict library for target ID
Predictive data analysis for CRISPR on-target activity

A pharmaceutical research team were designing CRISPR libraries with published scoring algorithms. However, they wanted to increase the efficiency of their libraries to improve data reproducibility of work being conducted in human A375 melanoma cells.


DESKGEN Predict Library
Parameter
Approach
Reference genome Human A375 cell line
Experimental intent Gene knockout
Nuclease SpCas9
Scoring algorithm Guide activity
44% improvement versus competing activity prediction

We determined a range of features that predict guide behavior in their specific cell line. We then trained a new model that outperformed the predictive ability of published scoring algorithms by 44%.

The new algorithm was integrated into DESKGEN AI and can be used to build unique libraries for this client in the future.

What can Desktop Genetics do for you?

I'm interested in