tSNE_project.R

#
#' @importFrom stats optim
#
project_map <- function(
  z,
  x_old,
  sum_X_old,
  x_old_tsne,
  P_tol = 5e-6,
  perplexity = 30
) {
  sum_z <- sum(z ^ 2)
  #x_old=test2@pca.rot[,1:5]
  #sum_X_old=rowSums((x_old^2))
  D_org <- sum_z + (-2 * as.matrix(x = x_old) %*% t(x = z) + sum_X_old)
  P <- d2p_cell(D = D_org, u = perplexity)
  nn_points <- which(x = P> P_tol) # Use only a small subset of points to comupute embedding. This keeps all the points that are proximal to the new point
  X_nn_set <- x_old[nn_points, ]  #Original points
  y_nn_set <- x_old_tsne[nn_points, ]  #Computed embeddings
  P_nn_set <- P[nn_points, ] #Probabilities
  y_new0 <- (
    t(x = as.matrix(x = y_nn_set)) %*%
      t(x = as.matrix(x = rbind(P_nn_set, P_nn_set)))
  )[, 1] #Initial guess of point as a weighted average
  sink("/dev/null")
  y_new <- optim(
    par = y_new0,
    fn = KullbackFun,
    gr = NULL,
    y_nn_set,
    P_nn_set,
    method = "Nelder-Mead"
  )
  sink()
  #plot(test2@tsne.rot)
  #points(y_new$par[1],y_new$par[2],col="red",cex=1,pch=16)
  #points(test2@tsne.rot[cell.num,1],test2@tsne.rot[cell.num,2],col="blue",cex=1,pch=16)
  #return(dist(as.matrix(rbind(y_new$par,test2@tsne.rot[cell.num,]))))
  return(y_new$par)
}

d2p_cell <- function(D, u = 15, tol = 1e-4) {
  betamin = -Inf
  betamax = Inf
  tries = 0
  tol = 1e-4
  beta = 1
  beta.list <- Hbeta(D = D, beta = beta)
  h <- beta.list[[1]]
  thisP <- beta.list[[2]]
  flagP <- beta.list[[3]]
  hdiff <- h - log(x = u)
  while (abs(x = hdiff) > tol && tries < 50) {
    if (hdiff > 0) {
      betamin <- beta
      if (betamax == Inf) {
        beta <- beta * 2
      } else {
        beta <- (beta + betamax) / 2
      }
    } else {
      betamax <- beta
      if (betamin == -Inf) {
        beta <- beta / 2
      } else {
        beta <- (beta + betamin) / 2
      }
    }
    beta.list <- Hbeta(D = D, beta = beta)
    h <- beta.list[[1]]
    thisP <- beta.list[[2]]
    flagP <- beta.list[[3]]
    hdiff <- h - log(x = u)
    tries <- tries + 1
  }
  # set the final row of p
  P <- thisP
  #Check if there are at least 10 points that are highly similar to the projected point
  return(P)
}

KullbackFun <- function(z, y, P) {
  #Computes the Kullback-Leibler divergence cost function for the embedding x in a tSNE map
  #%P = params{1};              %Transition probabilities in the original space. Nx1 vector
  #%y = params{2};              %tSNE embeddings of the training set. Nx2 vector
  print(z)
  print(dim(x = y))
  Cost0 = sum(P * log(x = P))      #Constant part of the cost function
  #Compute pairwise distances in embedding space
  sum_z <- sum(z ^ 2)
  sum_y <- rowSums(x = (y ^ 2))
  D_yz <- sum_z +(
    -2 * as.matrix(x = y) %*% t(x = matrix(data = z, nrow = 1)) + sum_y
  )
  Q <- 1 / (1 + D_yz)
  Q <- Q / sum(Q)               #Transition probabilities in the embedded space
  Cost <- Cost0 - sum(P * log(x = Q)) #% - 100 * sum(Q .* log(Q));
  return(Cost)
}

Hbeta <- function(D, beta) {
  flagP <- 1
  P <- exp(x = -D * beta)
  sumP <- sum(P)
  if (sumP < 1e-8) { #In this case it means that no point is proximal.
    P <- rep(length(x = P), 1) / length(x = P)
    sumP <- sum(P)
    flagP <- 0
  }
  H <- log(x = sumP) + beta * sum(D * P) / sumP
  P <- P / sumP
  return(list(H, P, flagP))
}