# Working with multiple trajectories

Very commonly, one has multiple trajectories of the same system, and we want to obtain the average results of all trajectories. We provide a simple scheme to average the results of multiple MDDF calculations:

### Create a vector of result data structures, without initialization

Let us assume that we have three Gromacs trajectories, with file names `traj1.xtc`

, `traj2.xtc`

, `traj3.xtc`

. First let us create a list with these file names:

`trajectory_files = [ "traj1.xtc" , "traj2.xtc" , "traj3.xtc" ]`

And define an empty vector of `Result`

structures:

`results = Result[]`

### Run the calculations in a loop

The calculation on the multiple trajectories is then performed in a simple loop, such as

```
atoms = PDBTools.readPDB("./system.pdb")
solute = AtomSelection(atoms,"protein",nmols=1)
solvent = AtomSelection(atoms,"resname TMAO",natomspermol=14)
for file in trajectory_files
trajectory = Trajectory(file,solute,solvent)
# compute the MDDF data and push the result to the results array
push!(results, mddf(trajectory))
end
```

### Merge the results of several trajectories, with proper weights

Of course, the resulting `results`

vector will contain at each position the results of each calculation. To merge these results in a single result data structure, use:

`R = merge(results)`

The `R`

structure generated contains the averaged results of all calculations, with weights proportional to the number of frames of each trajectory. That is, if the first trajectory had 2000 frames, and the second and third trajectories have 1000 frames each, the first trajectory will have a weight of 0.5 on the final results. The `merge`

function can be used to merge previously merged results with new results as well.

The names of the files and and weights are stored in the `R.files`

and `R.weights`

vectors of the results structure:

```
julia> R.files
3-element Array{String,1}:
"./traj1.xtc"
"./traj2.xtc"
"./traj3.xtc"
julia> R.weights
2-element Array{Float64,1}:
0.5
0.25
0.25
```

It is not a bad idea to check if that is what you were expecting.