Accurate Identification of SARS-CoV-2 from Viral Genome Sequences using Deep Learning

Alejandro Lopez-Rincon; Alberto Tonda; Lucero Mendoza-Maldonado; Eric Claassen; Johan Garssen; Aletta D. Kraneveld

CoV HKU-39849 [29] /SARS-CoV GDH-BJH01 organisms together, as they are all strains of SARS. Table 1 : Organism, assigned label, and number of samples in the unique sequences obtained from the repository [6] . We use the NCBI organism naming convention [30] . To encode the cDNA data into an input tensor for the CNN, we assigned numeric values to the different bases; C=0.25, T=0.50, G=0.75, A=1.0 (see Fig. 2 ). All missing entries were assigned the value 0.0. This procedure is 95 different from previous methods, that relied upon one-hot encoding [21, 20] , and has the advantages of making the input more human-readable and do not multiply the amount of memory required to store the information. We divide the available samples in two parts, 90% for training and validation (80% training, 10% validation), and 10% for testing, in a 10-fold cross-validation scheme. k- The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03.13.990242 doi: bioRxiv preprint The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.03.13.990242 doi: bioRxiv preprint