Nerve Block

In this study, we treated protein–ligand interaction prediction as a regression problem by aiming to predict the binding affinity scores. As a prediction model, we adopted a popular deep learning architecture, Convolutional Neural Network (CNN). CNN is an architecture that contains one or more convolutional layers often followed by a pooling layer. A pooling layer down-samples the output of the previous layer and provides a way of generalization of the features that are learned by the filters. On top of the convolutional and pooling layers, the model is completed with one or more fully connected (FC) layers. The most powerful feature of CNN models is their ability to capture the local dependencies with the help of filters. Therefore, the number and size of the filters in a CNN directly affects the type of features the model learns from the input. It is often reported that as the number of filters increases, the model becomes better at recognizing patterns (Kang et al., 2014 ).
We proposed a CNN-based prediction model that comprises two separate CNN blocks, each of which aims to learn representations from SMILES strings and protein sequences. For each CNN block, we used three consecutive 1D-convolutional layers with increasing number of filters. The second layer had double and the third convolutional layer had triple the number of filters in the first one. The convolutional layers were then followed by the max-pooling layer. The final features of the max-pooling layers were concatenated and fed into three FC layers, which we named as DeepDTA. We used 1024 nodes in the first two FC layers, each followed by a dropout layer of rate 0.1. Dropout is a regularization technique that is used to avoid over-fitting by setting the activation of some of the neurons to 0 (Srivastava et al., 2014 ). The third layer consisted of 512 nodes and was followed by the output layer. The proposed model that combines two CNN blocks is illustrated in Figure 2.
As the activation function, we used Rectified Linear Unit (ReLU) (Nair and Hinton, 2010 ), $g(x)=max(0,x)$ , which has been widely used in deep learning studies (LeCun et al., 2015 (link)). A learning model tries to minimize the difference between the expected (real) value and the prediction during training. Since we work on a regression task, we used mean squared error (MSE) as the loss function, in which P is the prediction vector, and Y corresponds to the vector of actual outputs. n indicates the number of samples.
$MSE=\frac{1}{n}{\displaystyle \sum _{i=1}^n{(P_i-Y_i)}^2}$
The learning was completed with 100 epochs and mini-batch size of 256 was used to update the weights of the network. Adam was used as the optimization algorithm to train the networks (Kingma and Ba, 2015 ) with the default learning rate of 0.001. We used Keras’ Embedding layer to represent characters with 128-dimensional dense vectors. The input for Davis dataset consisted of (85, 128) and (1200, 128) dimensional matrices for the compounds and proteins, respectively. We represented KIBA dataset with a (100, 128) dimensional matrix for the compounds and a (1000, 128) dimensional matrix for the proteins.

All procedures were approved by the UC Irvine Animal Care and Use Committee. Surgical procedures were performed stereotaxically under isofluorane anaesthesia and local nerve block induced by 0.5% bupivacaine. Kainic acid (50–100 nl, 20 mM in saline, Tocris Bioscience) was injected into the left dorsal hippocampus (2.0 mm posterior, 1.25 mm left, and 1.6 mm ventral to bregma) of mice on or after postnatal day 46. After recovery, animals were returned to the vivarium for at least 2 weeks to allow for the emergence of spontaneous recurrent seizures. Bipolar depth electrodes (PlasticsOne) and optical fibres (0.37NA, Low OH, 200 μm diameter, ThorLabs) terminated in 1.25 mm ceramic ferrules (Kientec Systems, Inc.) were implanted ipsilaterally (posterior 2.5 mm, left 1.75 mm, ventral 1.25 mm with respect to bregma) and in some cases, also contralaterally at the same posteroventral position into the hippocampus, targeting the dorsal stratum oriens of the CA1 such that emitted light would illuminate the hippocampal formation. Optical fibres and electrodes were fixed to the skull using screws (McMaster-Carr) and dental cement (Teets Cold Curing) and the animals were allowed to recover for several days before beginning 24-h video and EEG monitoring for seizures and subsequent closed-loop seizure detection and light delivery. On average, animals were implanted 15±2.3 weeks after KA injection and the effect of light on seizures was examined 15.9±1.4 weeks after KA injection (range: 2.4–24.6 weeks). There was no correlation between seizure duration reduction and time since KA for either Cam-HR or PV-ChR2 mice (P=0.39 and P=0.83, respectively, Spearman test; see also Supplementary Fig. S3).