Cancer diagnosis is critical in patient care yet it currently depends on time-consuming histopathology processes. We report a new method of computational staining in place of the traditional hematoxylin and eosin (H&E) staining. This method is derived from chemometric fluorescence microscopic imaging of unstained specimens. The computationally stained images visually differentiate specific cell properties, such as cellular metabolism of NADH, FAD, as well as protein production of tryptophan and elastin. Different color encoding strategies will be discussed including emulating the traditional H&E staining and optimizing for the contrast. The preliminary study on lung tissues suggests the proposed approach is a promising rapid histopathology alternative.