That's incorrect: open source software is necessarily available for redistribution. What you are referring to (source viewable but not necessarily available for redistribution/modification) is commonly referred as "source available" software [0].
Now, the distinction between open source and free software is more a social/political one rather than a practical one. Free software encompasses an entire philosophy while open source is simply concerned with software licensing.
The publication, yes. The structure, no. You can go download the structure of any and every known and published protein from any organism in the world for free with no copyright restrictions from the PDB. And that structure will be heavily scrutinized and reviewed before it's accessible.
That said, working in structural biology, it's incredible how quickly that field attached to open source, and in particular, free software. The big players in the field (coot, Relion, PEET and IMID) are all GPL. Additionally, software like PyMOL, Chimera, Phenix etc are available for free, sometimes with source. Those are basically all the software in the structural biologist toolset, with HKL2000 perhaps being the only paid software on anyone's workstation.
I suspect there's probably multiple reasons for this phenomenon. There's not many structural biologist, and of the few that there are, many are in academia and have no money to spend on software. Additionally, the field moves so rapidly that most software is actually piecemealed toolkits from tons of labs. Also, anything algorithmic requires substantial analysis if you want to publish data acquired with it, meaning it needs to be open. Finally, 99% of structural biology is done on Linux, so there's simply a culture aspect as well.
The data repository I mentioned is the PDB. Quoting http://www.rcsb.org/pages/policies :
> Data files contained in the PDB archive (ftp://ftp.wwpdb.org) are free of all copyright restrictions and made fully and freely available for both non-commercial and commercial use. Users of the data should attribute the original authors of that structural data.
That means it's available for redistribution.
BTW, the general trend is to use "source available" instead of "open source" when the source is available for viewing and perhaps local modification, but not for redistribution. https://en.wikipedia.org/wiki/Source-available_software
I think you misunderstand the distinction between open source and free software. Open source software also needs to be available for redistribution. The main practical difference is Free Software also requires that if you make any changes to the software, those changes must _also_ be released under the same licensing, whereas open source may or may not have such provisions.
See for instance this GNU page where permissive licenses like the MIT license are also described as "free software licenses":
"The term “open source” software is used by some people to mean more or less the same category as free software. It is not exactly the same class of software: they accept some licences that we consider too restrictive, and there are free software licences they have not accepted. However, the differences in extension of the category are small: nearly all free software is open source, and nearly all open source software is free."
In particular, he notes the example of DRM software. DRM software can be distributed under an open-source licence, but by design and in practice it serves to restrict the freedoms of its users, not to enhance them.
While Stallman says you can charge as much as you can, his terms say that once someone buys it they can immediately undercut you, and he admits that limits the amount you could charge sustainably.