On the use of the scheraga–mandelkern equation in determining molecular weights of very asymmetric proteins