spateo.tools.spatial_impute.impute_model

Module Contents

Classes

Discriminator	Module that learns associations between graph embeddings and their positively-labeled augmentations
AvgReadout	Aggregates graph embedding information over graph neighborhoods to obtain global representation of the graph
Encoder	Representation learning for spatial transcriptomics data

class spateo.tools.spatial_impute.impute_model.Discriminator(nf: int)

Bases: torch.nn.Module

Module that learns associations between graph embeddings and their positively-labeled augmentations

Parameters

nf

Dimensionality (along the feature axis) of the input array

weights_init(m)

forward(g_repr: torch.FloatTensor, g_pos: torch.FloatTensor, g_neg: torch.FloatTensor)

Feeds data forward through network and computes graph representations

Parameters

g_repr

Representation of source graph, with aggregated neighborhood representations

g_pos

Representation of augmentation of the source graph that can be considered a positive pairing, with aggregated neighborhood representations

g_neg

Representation of augmentation of the source graph that can be considered a negative pairing, with aggregated neighborhood representations

Returns

Similarity score for the positive and negative paired graphs

Return type

logits

class spateo.tools.spatial_impute.impute_model.AvgReadout

Bases: torch.nn.Module

Aggregates graph embedding information over graph neighborhoods to obtain global representation of the graph

forward(emb: torch.FloatTensor, mask: torch.FloatTensor)

Parameters

emb

float tensor Graph embedding

mask

float tensor Selects elements to aggregate for each row

class spateo.tools.spatial_impute.impute_model.Encoder(in_features: int, out_features: int, graph_neigh: torch.FloatTensor, dropout: float = 0.0, act=F.relu, clip: Union[None, float] = None)

Bases: torch.nn.modules.module.Module

Representation learning for spatial transcriptomics data

Parameters

in_features

Number of features in the dataset

out_features

Size of the desired encoding

graph_neigh

Pairwise adjacency matrix indicating which spots are neighbors of which other spots

dropout

Proportion of weights in each layer to set to 0

act

object of class torch.nn.functional, default F.relu. Activation function for each encoder layer

clip

Threshold below which imputed feature values will be set to 0, as a percentile of the max value

reset_parameters()

forward(feat: torch.FloatTensor, feat_a: torch.FloatTensor, adj: torch.FloatTensor)

Parameters

feat

Counts matrix

feat_a

Counts matrix following permutation and augmentation

adj

Pairwise distance matrix