Figure 1
From: PACIFIC: a lightweight deep-learning classifier of SARS-CoV-2 and co-infecting RNA viruses
Overview of PACIFIC and its model architecture (a) PACIFIC uses FASTQ or FASTA files as inputs to make read-level predictions and to estimate the relative number of reads belonging to each RNA virus or human class. (b) PACIFIC deep neuronal network architecture, which includes embedding of k-mer tokens, convolutional neural network filtering to BiLSTM layers. The model is trained using in silico generated sequences from RNA virus genomes and the human transcriptome. CNN—convolutional neural network; BiLSTM—bi-directional long short-term memory.
