Bin distribution. The plot to the left (A) shows the distribution of the 20 most populated 30° × 30° bins for all coil residues in the training set. The plot in the middle (B) shows the distribution for just threonines in the training set, and the plot to the right (C) shows the predicted bins for threonine in the sequence Glu-Leu-Asp-Thr-Glu-Asp-Ala taken from a randomly chosen protein in the data set. The neighboring residues are used by the neural network to suggest a different distribution to yield a higher success rate. The darker the color of the bin the more likely it is that the angle set is within this bin.