1: \begin{abstract}

2: By assigning a probability measure via the spectrum of the normalized Laplacian to each graph and

3: %For each graph, we associate the spectrum of the normalized

4: %Laplacian with a probability measure, called the spectral measure.

5: using $L^p$ Wasserstein distances between probability measures, we

6: define the corresponding spectral distances $d_p$ on the set of all

7: graphs. This approach can even be extended to measuring the distances between infinite graphs. We prove that the diameter of the set of graphs,

8: as a pseudo-metric space equipped with $d_1$,

9: %$L^1$ Wasserstein distance

10: is one. We further study the behavior of $d_1$ when the size of graphs tends to infinity by interlacing inequalities aiming at exploring large real networks.

11: A monotonic relation between $d_1$ and the evolutionary distance of biological networks is observed in simulations.

12: %Furthermore, by the interlacing results of the spectra we study

13: %the convergence of graphs under the spectral distances. We also

14: %explore some applications in biology.

15: 

16: \smallskip

17: \noindent \textsc{Keywords.} Wasserstein distance, spectral measure, random rooted graph, asymptotic behavior, biological networks

18: 

19: 

20: \smallskip

21: \noindent \textsc{AMS subject classifications.} 05C50, 28A33, 05C63, 92B10

22: \end{abstract}

23: