The scLDVAE model

mutable struct scLinearDecoder <: AbstractDecoder

Julia implementation of a linear decoder for a single-cell LDVAE model corresponding to the scvi-tools linear decoder Collects all information on the decoder parameters and stores the decoder parts. Can be constructed using keywords.

Keyword arguments

• n_input: input dimension = dimension of latent space
• n_output: output dimension of the decoder = number of genes/features
• factor_regressor: Flux.Chain of fully connected layer + optional normalisation realising the first part of the decoder (before the split in mean, dispersion and dropout decoder). For details, see the source code of FC_layers in src/Utils. Only one layer without activation.
• px_dropout_decoder: if the generative distribution is zero-inflated negative binomial (gene_likelihood = :zinb in the scVAE model construction): Flux.Dense layer, else nothing.
• px_r_decoder: decoder for the dispersion parameter. If generative distribution is not some (zero-inflated) negative binomial, it is nothing. Else, it is a parameter vector or a Flux.Dense, depending on whether the dispersion is estimated per gene (dispersion = :gene), or per gene and cell (dispersion = :gene_cell)
• use_batch_norm: whether or not to apply batch normalization in the decoder layers
• use_layer_norm: whether or not to apply layer normalization in the decoder layers

scLinearDecoder(n_input, n_output; 
    bias::Bool=true,
    dispersion::Symbol=:gene,
    gene_likelihood::Symbol=:zinb,
    dropout_rate::Float32=0.0f0,
    use_batch_norm::Bool=true,
    use_layer_norm::Bool=false
    )

Constructor for a linear decoder for an scLDVAE model. Initialises a scLinearDecoder struct with the parameters specified by the inputs. Julia implementation of the scvi-tools linear decoder.

Arguments:

• n_input: number of input features for the decoder; has to be equal to the latent space dimension.
• n_output: number of features in the final output layer of the decoder, has to be equal to the number of genes in the dataset.

Keyword arguments:

• bias: whether or not to use bias parameters in the neural network layers
• dispersion: can be either :gene or :gene-cell. The Python scvi-tools options :gene-batch and gene-label are planned, but not supported yet.
• dropout_rate: Dropout to use in the encoder and decoder layers. Setting the rate to 0.0 corresponds to no dropout.
• gene_likelihood: which generative distribution to parameterize in the decoder. Can be one of :nb (negative binomial), :zinb (zero-inflated negative binomial), or :poisson (Poisson).
• use_batch_norm: whether or not to apply batch normalization in the encoder/decoder layers
• use_layer_norm: whether or not to apply layer normalization in the encoder/decoder layers

scLDVAE(n_input::Int;
    activation_fn::Function=relu, # to be used in all FC_layers instances
    bias::Symbol=:both,  # :both, :none, :encoder, :decoder; whether to use bias in linear layers of all FC instances in encoder/decoder
    dispersion::Symbol=:gene,
    dropout_rate::Float32=0.1f0,
    gene_likelihood::Symbol=:zinb,
    latent_distribution::Symbol=:normal,
    library_log_means=nothing,
    library_log_vars=nothing,
    log_variational::Bool=true,
    n_batch::Int=1,
    n_hidden::Int=128,
    n_latent::Int=10,
    n_layers::Int=1,
    use_activation::Symbol=:both, # :both, :none, :encoder, :decoder
    use_batch_norm::Symbol=:both, # :both, :none, :encoder, :decoder
    use_layer_norm::Symbol=:none, # :both, :none, :encoder, :decoder
    use_observed_lib_size::Bool=true,
    var_activation=nothing,
    var_eps::Float32=Float32(1e-4),
    seed::Int=1234
)

Constructor for a linearly decoded VAE model. Initialises an scVAE model with a linear decoder with the parameters specified in the input arguments. Julia implementation of the scvi-tools LDVAE object. Differs from the scVAE constructor only in that it defines a linear decoder, see scLinearDecoder.

Arguments:

• n_input: input dimension = number of genes/features

Keyword arguments

• activation_fn: function to use as activation in all neural network layers of encoder and decoder
• bias: whether or not to use bias parameters in the neural network layers of encoder and decoder
• dispersion: can be either :gene or :gene-cell. The Python scvi-tools options :gene-batch and gene-label are planned, but not supported yet.
• dropout_rate: Dropout to use in the encoder and decoder layers. Setting the rate to 0.0 corresponds to no dropout.
• gene_likelihood: which generative distribution to parameterize in the decoder. Can be one of :nb (negative binomial), :zinb (zero-inflated negative binomial), or :poisson (Poisson).
• library_log_means: log-transformed means of library size; has to be provided when not using observed library size, but encoding it
• library_log_vars: log-transformed variances of library size; has to be provided when not using observed library size, but encoding it
• log_variational: whether or not to log-transform the input data in the encoder (for numerical stability)
• n_batch: number of batches in the data
• n_hidden: number of hidden units to use in each hidden layer
• n_latent: dimension of latent space
• n_layers: number of hidden layers in encoder and decoder
• use_activation: whether or not to use an activation function in the neural network layers of encoder and decoder; if false, overrides choice in actication_fn
• use_batch_norm: whether to apply batch normalization in the encoder/decoder layers; can be one of :encoder, :decoder, both, :none
• use_layer_norm: whether to apply layer normalization in the encoder/decoder layers; can be one of :encoder, :decoder, both, :none
• use_observed_lib_size: whether or not to use the observed library size (if false, library size is calculated by a dedicated encoder)
• var_activation: whether or not to use an activation function for the variance layer in the encoder
• var_eps: numerical stability constant to add to the variance in the reparameterisation of the latent representation
• seed: random seed to use for initialization of model parameters; to ensure reproducibility.