1: \begin{abstract}

2: We consider a chemical reaction network governed by mass action kinetics

3: and composed of $N$ different species which can reversibly form heterodimers.

4: A fast iterative algorithm is introduced to compute the equilibrium

5: concentrations of such networks. We show that the convergence is

6: guaranteed by the Banach fixed point theorem. As a practical example, of

7: relevance for a quantitative analysis of microarray data, we consider

8: a reaction network  formed by $N \sim 10^6$ mutually hybridizing

9: different mRNA sequences. We show that, despite the large number of

10: species involved, the convergence to equilibrium is very rapid for

11: most species.

12: The origin of slow convergence for some specific subnetworks is discussed. 

13: This provides some insights for improving the performance of the algorithm.

14: \end{abstract}

15: