Skip to main content
Fig. 1 | BMC Bioinformatics

Fig. 1

From: DEGnext: classification of differentially expressed genes from RNA-seq data using a convolutional neural network with transfer learning

Fig. 1

Workflow of DEGnext methodology. In this workflow, there are three main phases. The first phase involves data collection, preprocessing, labelling, and splitting of the data. Here, we split the data into two parts: non-biologically validated data (“non-bio data” or P) and the biologically validated data (“bio data” or Q). T1 is the non-biologically validated train data of P (“non-bio train data”, 80% of P). T2 is the non-biologically validated test data of P (“non-bio test data”, 20% of P). F1 is the fine-tune data of biologically validated data (“fine-tune data”, 80% of Q). T3 is the biologically validated test data of Q (“bio-test data”, 20% of Q). The second phase includes training (first level training) and fine-tuning (second level training) and testing of CNN model to predict UR and DR genes. The third phase includes downstream enrichment analyses of the predicted UR and DR genes to identify potential biomarkers related to a cancer dataset. The CNN architecture is illustrated in Fig. 2

Back to article page