ML-Pred-CLL: Machine Learning based prediction of Chronic Lymphocytic Leukemia using protein sequential data

Leukemia is a fatal disease caused by mutation in the blood forming tissues in the bone marrow. Leukemia is the deadliest disease pertaining to cancer. Early detection of leukemia is necessary to increase the chances of survival. In this work, we have proposed Machine Learning based prediction of Chronic Lymphocytic Leukemia (CLL) using protein sequential data (ML-Pred-CLL). In this study, the most frequently mutated genes of CLL - BRAF, MYD88, NOTCHI, and XPOI - were used. The significant features of Chou’s Pseudo Amino Acid Composition (PseAAC) and relative/position-based features in protein sequences contributed towards the development of effective chronic lymphocytic leukemia classification model. Several machine learning models i.e. Support Vector Machine(SVM), XGBoost, Random Forest, K Nearest Neighbor(KNN), Logistic Regression, and Naive Bayes, have reported 97.09%, 95.17%, 93.5%, 91.30%, 70.1%, 54.03% accuracy respectively. The models were trained and tested with stratified k-fold cross validation technique. The primary benefit of the suggested ML-Pred-CLL is that it effectively utilizes the discriminant information of mutant genes. Using discriminant knowledge from mutant genes, we explored at a new dimension of early chronic lymphocytic leukemia classification. The proposed ML-Pred-CLL is highly effective for early diagnosis and prognosis of chronic lymphocytic leukemia. Additionally, the suggested ML-Pred-CLL is expected to be a helpful tool for oncologists.

I. Introduction

The number of Leukemia related deaths is rapidly increasing around the world. It is caused by mutations in the blood forming tissues in the bone marrow [1]. Leukemia affects the lifecycle of white blood cells in the body. The kind of white blood cells involved and the rate of illness progression are used to categorize it. Leukemia is divided into acute and chronic forms based on the disease’s progression. In acute leukemia, the white blood cells divide uncontrollably and do not mature the way they should. In chronic leukemia the white blood cell matures partly but live longer than normal white blood cells leading to build up in organs [2]. Based on which bone marrow cell the cancer starts in, leukemia is either myeloid or lymphocytic. This gives rise to four different forms of Leukemia, i.e. Chronic lymphocytic leukemia, acute myeloid leukemia, acute lymphocytic leukemia, and chronic myeloid leukemia (CML) [2].