Will open science change chemistry? | Hacker News Reader

Will open science change chemistry?

chemistryworld.com

78 pointsdaphnemichala26 comments

This article conflates a lot of different “open” ideas. Open access journals, machine readable datasets, standardized data representation, open community engagement, etc. Each of which is challenging on its own, let alone hoping to tackle all at once.

Chemistry has the blessing and the curse of being an older disciple. Fundamentals have not changed in decades. If you grab an organic chemistry text from the 70s, I guess you would be missing on some cutting edge reactions, and you would be using hilariously obsolete analytical techniques, but the synthesis is going to be the same.

Which is to say, that I do not believe chemistry is particularly held back at this point. I think the author was angling for, “Would it not be great if we had standardized, digitized chemistry reaction libraries so we could have an AlphaFold moment?” Which sure, but someone is going to have to fund the effort of digitizing decades of chemistry knowledge.

We already have standardized, digital chemistry reaction libraries (reaxys, scifinder). They cost money, but are heavily used by organic chemists.

More the focus on open libraries. I am not aware of anything of significant size, but I have not been in the space for a while.

There are attempts.

https://open-reaction-database.org

Is there any open-source structure-based search?

Sure. RDKit supports SMARTS.

https://www.rdkit.org/docs/source/rdkit.Chem.rdchem.html#rdk...

Interesting. Is there a chemical database you can plug this into?

Use any database you wish. They all use standard formats, which RDKit understands.

I've been playing with Avogadro lately, and some of the libraries seem pretty advanced.

You know it's legitimate science when Fortran libraries start compiling as part of the dependencies.

https://avogadro.cc/

> Twenty years ago the debate surrounding open science focused on access to journals. By 2020 around 25% of all chemistry papers published were open access, and now most of the major publishers of chemistry journals offer some version of open access.

I'm a big fan of open source and open access, but I'm not sure that access to journals is really a big problem for any working chemist. So if the question is whether it will "change chemistry", I would say it's unlikely.

Why? This isn't an easy field and it's full of landmines, sometimes literally so. It's really only possible to work in the field if you're in a well-funded lab in a well-funded university. Not only that, but many of these chemicals are dangerous. You're just not supposed to be dorking around with them in your basement or garage.

It's nice to imagine that somewhere out there is some poor, underfunded genius who doesn't have $100 or $200 to pay for a copy of some article, but I think from a practical perspective that's just not something holding back the field. And really, most of his or her neighbors would be happy if he/she wasn't working with dangerous chemicals in a garage.

Yep, as much as I love the idea, I would not recommend that anyone approaches even some of the basic chemistry without someone being able to stop you in time.

I’m pretty sure the majority of chemistry teachers I’ve had with real research or professional experience over the age of 50 were blind or deaf on at least one side. The remainder have a story about how they almost died.

>It's really only possible to work in the field if you're in a well-funded lab in a well-funded university.

It takes a lot more ambition than most people have, to do it otherwise, and it's now more difficult than ever, but it's been done.

>many of these chemicals are dangerous.

The first laboratory I built on my own was in a room rated by the fire marshal for storage of hazardous materials. So we could work with many kilos of chemicals and samples without running afoul of the regulations. I personally made sure there was way better ventilation than they have in the chemical plants before instrumenting the lab.

There were neighbors but it was an industrial suburb. For a while there they had more welding tanks, drums of paint, heavy diesel trucks, etc than we did toxins but eventually we outdid them with our growing storage sheds out in the yard.

You just don't try this at home.

You don't realize how nice it is to have a refuge from toxicity when that's not what you have at work.

I'm a strong proponent of open data but I think it needs to go beyond science. Companies should also be incentivized or even required for keeping designs, data, interfaces etc. open. Because often scientists are reliant on devices and data platforms that restrict ownership. So even if the scientist wants to keep his data open, a product or a service that the scientist uses for his research might prevent this.

> Companies should also be incentivized or even required

required sounds bad because you make the costs higher and prevents small companies from entering the market once its too much. as for incentives any idea on how that could work?

Marginally lower corporate tax rate.

After the floor on corporate tax rate is returned to a sustainable level.

I don’t understand the use of the word floor here. The current US corp tax is 21%. The average corporate tax rate in the EU is 21.3%. Are you saying the corporate tax rate in the EU should be higher?

I remember a time where high school teachers thought it was a good idea to do a demo of iodine mixed with ammonia to get NI3 which goes boom when touched and produces this nice purple cloud...to shut up the classroom students....

My Chemistry teach was a weird bass...he also taught diving to the local police departments for lake rescue in Lake Michigan and rode a motorcycle to school.

https://github.com/FourThievesVinegar/solderless-microlab

> The MicroLab is an open-source, DIY, automated controlled lab reactor (CLR) that people can assemble with parts available online. We hope this will do for chemistry what the 3D printer did for manufacturing.

I think it’s okay if people take responsibility for their actions, to take risks known or unknown. It’s okay to live in a world where we’re not entirely insulated from negative consequences. Ive always been bummed by how inaccessible chemistry is.

Back when I ran a maker-space in Seattle we interviewed this dude who had a full bio-safety rated lab in his garage. I wish to see more of that stuff!

SoundBio is a bio + chemistry makerspace in Seattle. Not BSL, though.

As a person who grew up with:

https://www.goodreads.com/book/show/2478325.The_Golden_Book_...

this is mostly a societal thing --- if one wants to do experiments, it's pretty straight-forward to research and begin and learn.

Unfortunately computational quantum chemistry still dominated by proprietary software, like the infamous Gaussian. It seems physicists have better open source tools and culture,by comparison.

Gaussian's main advantage is the GUI!

If you need a gaussian basis set code, nwchem exists! Orca is another one.

If you need a plane wave code but don't want to shell for VASP, Quantum Espresso is awesome, and easy to compile.

Then there's CP2k, whose gaussian basis with planewave auxiliaries are pretty decent at molecular dynamics boxes.

And to set up structures in a GUI, Avogadro is very useful for molecules. VESTA is great at crystals if you are willing to do a little math. And in Python, ASE and Pymatgen will do everything else that the GUI can't.

Don't get me wrong, Gaussian is a fantastic program, and their developers deserve every penny - but there's no scientific reason you can't use free codes!

IMO, the big limiter to open science in computational chemistry isn't the codes. It's the lack of computer power the average consumer has.

Many of the DFT calculations I needed for my PhD thesis were done in NWChem. Out of spite? Not really, but it did make me feel good to use something non-proprietary.

Reply on news.ycombinator.com