torchdms.plot

Plotting functions.

Functions

beta_coefficients	This function takes in a (ideally trained) model and plots the values of the weights corresponding to the inputs, dubbed "beta coefficients".
build_2d_nonlinearity_df	Build a dataframe that contains the value of the nonlinearity for the domain of the range of values seen in the geplot_df.
build_geplot_df	Build data frame for making a global epistasis plot.
df_with_named_columns_of_np_array
plot_2d_geplot
plot_error
plot_exploded_binarymap_dataframe_summary
plot_geplot
plot_heatmap	This function takes in a model and plots the single mutant predictions.
plot_observed_2d_scores	If a dataset has two targets, create a scatterplot of the scores.
plot_svd	This function plots the log singular values and the cummulative sum of each of a trained model's beta coefficent matricies.
plot_svd_profiles	Plots heatmaps of amino acid profiles (U) and site profiles (V) from SVD output after final gradient step of training.
plot_test_correlation	Plot scatter plot and correlation values between predicted and observed for each target.
pretty_breaks
series_min_max

torchdms.plot.plot_test_correlation(evaluation_dict, model, out, cmap='plasma')

Plot scatter plot and correlation values between predicted and observed for each target.

torchdms.plot.plot_heatmap(model, path)

This function takes in a model and plots the single mutant predictions.

We plot this as a heatmap where the rows are substitutions nucleotides, and the columns are the sequence positions for each mutation.

torchdms.plot.beta_coefficients(model, test_data, out)

This function takes in a (ideally trained) model and plots the values of the weights corresponding to the inputs, dubbed “beta coefficients”.

We plot this as a heatmap where the rows are substitutions nucleotides, and the columns are the sequence positions for each mutation.

torchdms.plot.build_geplot_df(model, data, device='cpu')

Build data frame for making a global epistasis plot.

torchdms.plot.build_2d_nonlinearity_df(model, geplot_df, steps)

Build a dataframe that contains the value of the nonlinearity for the domain of the range of values seen in the geplot_df.

torchdms.plot.plot_svd(model, test_data, out)

This function plots the log singular values and the cummulative sum of each of a trained model’s beta coefficent matricies.

torchdms.plot.plot_svd_profiles(model, test_data, out)

Plots heatmaps of amino acid profiles (U) and site profiles (V) from SVD output after final gradient step of training.

torchdms.plot.plot_observed_2d_scores(data, targets, out)

If a dataset has two targets, create a scatterplot of the scores.