Myosin VI is a motor protein that attaches to a suitable substrate called Actin filament. This particular protein travels towards the minus end of Actin filament, which is in an opposite direction as compared to other proteins in the Myosin family. To study the dynamics of Myosin VI proteins, we may be able to find the answers of biological questions related to their roles in intracellular transport. In order to study the dynamics of such protein, we need to capture their spatio-temporal movements using microscopy imaging techniques. To this end, Total Internal Reflection Fluorescence Microscopy (TIRFM) has recently become the most popular one among such imaging techniques. A single TIRFM sequence comprises of thousands of fluorescent protein spots over several hundreds of frames. Manual tracking in large TIRFM datasets suffers from unrepeatability, inefficiency, and inaccuracy due to subjective assessment. Thus, a computer-aided tracking (CAT) system is well sought after by cell biologists. However, automated tracking is faced with many challenges in noisy and senesce TIRFM sequences, i.e., a large population of spots, frequent and persistent spot overlapping, appearing and disappearing, and abruptly changing of spot direction and speed. In this paper, we proposed a novel automated spot tracking framework, which overcomes these challenges raised in noisy and dense TIRFM sequences. The experimental results show that our framework achieves higher tracking accuracies compared to the state-of-the-art tracking methods, especially for sequences with high spot density.