Table 7 CarboxE site identification model based on CNN.

From: Computational identification of 4-carboxyglutamate sites to supplement physiological studies using deep learning

Layer type

No. of weights

Embedding layer to convert numeric sequence into vector sequence

\((23 \times 20 = 460)\)

Conv-maxpool-1D block with 10 filters of size 5

\(((5 \times 20) + 1) \times 10 = 1010\)

Dropout with 25% of probability

N/A

Conv-maxpool-1D block with 16 filters of size 3

\(((3 \times 10) + 1) \times 16) = 496\)

GlobalAveragePooling1D

N/A

Dropout with 50% of probability

N/A

Dense layer with 8 units

\((16+1) \times 8 = 136\)

Output layer

\((8+1) \times 1 = 9\)

Back to article page