I am assuming you performed a 10-class heterogeneous refinement to explore heterogeneity in the dataset. If this is the case, you should examine the volumes visually and inspect the orientation distribution plots for each class as well as the number of particles, to inform whether there are in fact distinct conformations/states/etc.
To obtain a single volume from all the particles, you can run an Ab-Initio Reconstruction job using the default 1 class, and then take all the particles from the Ab-Initio Reconstruction along with the volume that was generated, and put those into a Homogeneous Refinement (NEW) job to generate a single refined structure.
Please let us know if there are other specific considerations. Thanks!