Usage

To use sgRNA Modeler in a project:

import sgrna_modeler

class sgrna_modeler.models.SklearnSgrnaModel(random_state=7, val_frac=0.1, model=None, features=None)

scikit-learn gradient boosting for modeling sgRNA activity

Parameters:

• random_state (int) – set random state in train/test split for reproducibility
• val_frac (float) – amount of data to use for early stopping
• model (sklearn GradientBoostingRegressor) – base model
• features (list) – features to model

Example:

>>> from sgrna_modeler import datasets as da
>>> from sgrna_modeler import models as sg
>>> train_model = sg.SklearnSgrnaModel()
>>> rs2_data = da.load_doench_2016()
>>> train_model.fit(rs2_data)

sgrna_modeler.models.build_kim2018(input_shape=(34, 4))

Build a convolutional neural network

From: Kim, Hui Kwon, et al. “Deep learning improves prediction of CRISPR–Cpf1 guide RNA activity.” Nature biotechnology 36.3 (2018): 239.

Parameters:input_shape (tuple) – guide length by nts (4) Returns:CNN architecture Return type:keras Model object