Title | ||
---|---|---|
Down syndrome prediction/screening model based on deep learning and illumina genotyping array |
Abstract | ||
---|---|---|
Down syndrome (DS) is a genetic disorder with genome dosage imbalances and micro-duplications of human chromosome 21. It is usually associated with a group of serious diseases, including intellectual disabilities, cardiac diseases, physical abnormalities, and other abnormalities. Currently, since there is no cure for human DS, screening and early detection have become the most efficient way for DS prevention. In this study, we used deep learning techniques to build accurate DS prediction/screening models based on the analysis of newly introduced Illumina genotyping array. Specifically, we built chromosome SNP maps based on clinical genotyping data collected by Vanderbilt University Medical Center. Then we proposed a convolutional neural network (CNN) architecture with ten layers and two merged CNN models, which took two input chromosome SNP maps in combination. Our CNN DS prediction/screening model achieved over 99.3% average accuracy, as well as very low false positive and false negative rate, which are critical to disease prediction and screening in medical practice. It also had better performances in terms of all evaluating metrics when compared with three conventional machine-learning algorithms. Finally, we visualized the feature maps and the trained filter weights from intermediate layers of our trained CNN model. We further discussed the advantages of our method and the underlying reasons for its robust performance. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/BIBM.2017.8217674 | 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) |
Keywords | DocType | ISBN |
Deep Learning,Convolutional Neural Network,Down Syndrome,Illumina Genotyping | Conference | 978-1-5090-3051-4 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
7 |
Name | Order | Citations | PageRank |
---|---|---|---|
Bing Feng | 1 | 0 | 0.34 |
David C. Samuels | 2 | 0 | 0.34 |
William Hoskins | 3 | 0 | 0.34 |
Yan Guo | 4 | 0 | 0.34 |
Yan Zhang | 5 | 133 | 30.68 |
Jijun Tang | 6 | 370 | 48.23 |
Zibo Meng | 7 | 248 | 13.60 |