TY  - JOUR
TI  - Convolutional Neural Network Approach to Predict Tumor Samples Using Gene Expression Data
AB  - Cancer is threatening millions of people each year and its early diagnosis is still a challenging task. Early diagnosis is one of the major ways to tackle the disease and lower the mortality rate. Advancements in deep learning approaches and the availability of biological data offer applications that can facilitate the diagnosis and characterization of cancer. Here, we aimed to provide a new perspective of cancer diagnosis using a deep learning approach on gene expression data. In this study, RNA-Seq data of approximately 30 different types of cancer patients the Cancer Genome Atlas (TCGA) study, and normal tissue RNA-Seq data from GTEx were used. The input data for the training was transformed to RGB format and the training was carried out with a Convolutional Neural Network (CNN). The trained algorithm is able to predict cancer with 97% accuracy, using gene expression data. In conclusion, our study shows that the deep learning approach and biological data have a huge potential in the diagnosis and identification of tumor samples.
AU  - Darendeli Kiraz, Büşra Nur
AU  - Yilmaz, Alper
DO  - 10.38016/jista.946954
PY  - 2021
JO  - Zeki sistemler teori ve uygulamaları dergisi (Online)
VL  - 4
IS  - 2
SN  - 2651-3927
SP  - 136
EP  - 141
DB  - TRDizin
UR  - http://search/yayin/detay/494815
ER  -