Проф. др Владимир Брусић, почасни члан Српског Кривака, одржаће предавање под називом “CLASSIFICATION OF FIVE CELL TYPES FROM PBMC SAMPLES USING SINGLE CELL TRANSCRIPTOMICS AND ARTIFICIAL NEURAL NETWORKS” у Математичком институту САНУ у уторак 10.03.2020. u 14:15, сала 301f.
We used 27 human single cell transcriptomics (SCT) data sets to develop an artificial neural network (ANN) model for classification of Peripheral Blood Mononuclear Cells (PBMC). We demonstrated that highly accurate models for classification of PBMC subtypes can be developed by combining multiple independent data sets to form training data sets. A significant data preparation effort was needed for building predictive models. Using a data set of ~120,000 single cell instances we showed the accuracy of classification of PBMC call of ~90%. Optimization techniques and addition of new high-quality data sets for model training are expected to improve PBMC subtype classification accuracy. The exclusion of genes that have zero transcript expression and the addition of new files with 32 cycles of training and testing has improved classification accuracy to ~92%.