Contributions to Digital Health in Low and High Level
14:00 - 14:20
Building 19, Hall 1
Digital health (DH) is an important topic these days since it provides a lot of help to the smart world that we are approaching. This study contributes to digital health on different levels by introducing methods based on machine learning (ML).
The first contribution concerns developing generic models for the recognition of different genomic signals and regions (GSR) within eukaryotic DNA, which is a low level in DH. We developed DeepGSR, a systematic framework that facilitates generating ML models to predict GSR with high accuracy. To the best of our knowledge, no available generic and automated method exists for such task that could facilitate the studies of newly sequenced organisms and detect some diseases that are caused by specific genes. The prediction module of DeepGSR uses deep learning algorithms to derive highly abstract features that depend mainly on proper data representation and hyperparameters calibration. DeepGSR, which was evaluated on recognition of PAS and translation initiation sites (TIS) in different organisms, yields a more straightforward and more precise representation of the problem under study, compared to some other hand-tailored models, while producing high accuracy prediction results.
Secondly, we focus on deriving a model capable of facilitating the early detection of breast cancer, which is a high level in DH. We provide E-Detect, a full system with frontend (website) and backend (deep learning model), that allows the users to upload the mammograms and collect the results of the probability of having breast cancer. E-Detect uses both image processing and deep learning models and produces an acceptable prediction accuracy.
Overall, our research contributed to different aspects of DH.