One of the diseases with a sharply increasing incidence rate is bowel cancer. Intestinal cancer is an internal disease that frequently advances without obvious signs, making detection and identification difficult. Nonetheless, early diagnosis is essential because it improves patient outcomes by lowering diagnostic delays, easing the clinical load, and facilitating more efficient treatment approaches. Using a variety of data sets, including histopathology pictures, several deep learning models have been created in recent years to help doctors diagnose bowel cancer early. In this work, we present a new method in this work for identifying intestinal cancer patients and differentiating them from tumor suppressor gene-related non-pathological lesions. We use two different datasets and rely on a deep learning system. Protein sequences from the APC gene that were obtained through SNPs make up the first dataset. This gene is located at q22.2 on human chromosome 5. The second set of data contains the same types of sequences with the difference of being located on the TP53 gene at chromosome 17. The SNP sequences of the two data sets are transformed from their textual forms into scalogram images under different colorimetric spaces. For classification, we opted for the Convolutional Neural Network (CNN). Various quantitative metrics, including precision, sensitivity, specificity, and accuracy, are then employed for evaluation. The TP53 gene scalograms were classified with the highest accuracy, sensitivity, specificity, and precision rates of 96.37, 98.83, 97.41, and 60.00%, respectively. As for the APC gene, an accuracy value of 96.19%, a sensitivity of 97.42%, a specificity of 85.46%, and a precision of 98.32% were recorded. These results show the robustness of the combination of our CNN architecture and the tumor suppressor APC and TP53 genes in terms of early intestinal cancer identification.