Principal component analysis based unsupervised feature extraction applied to budding yeast temporally periodic gene expression

What is it about?

In order to find which genes are temporally periodic, you need to know not only period but also functional forms. Since it is usually impossible to know the periodic functional form a priori, although the sinusoidal function is usually assumed, it is not guaranteed to be suitable one. In this paper, we demonstrated that PCA based unsupervised FE can solve this difficulty, i.e., we can identify periodic genes without knowledge of either period or functional form.

Featured Image

Why is it important?

We find that gene expression often does not follow sinusoidal function, Thus, it is important to know which one is periodic without specifying functional forms. Our methodology, PCA based unsupervised FE enables us to do this task.

Perspectives