Support chained experiments linked by common samples

[jmmitc06]

As discussed with Stephen Barnes.

For example xpresss and mpress samples.

[shuzhao-li]

I should include this in the implementation of join subcommand.

[jmmitc06]

In essence this is achieved by selecting references manually within each batch prior to joining. This could be made more explicit in the future.

[jmmitc06]

With [implement_join 3dc7ebe] pulled, this is comes down to selecting the appropriate references manually while peak picking each experiment. Will close.