Inverse Design – A Path to Faster Development?

First in a continuum of brief news reports on Brown University’s Chemistry Department Seminar series. Prof. Dequan Xiao, from the University of New Haven’s Center for Integrative Materials Discovery, presented applications of his team’s “inverse” design methodology, and explained its future impacts.

I find myself stymied by the mountain of choices set before me, threatening to tumble down any moment. I know what I came for though: a shade of dark green and minimal abrasions, promising the perfect composition. My eyes scan the evergreen cliff face, slowly narrowing my search to a few prime options near the summit as I toss away paths that lead me elsewhere, until… bingo.

A perfect avocado is mine.

Material design currently relies on brute force searching, otherwise known as guess-and-check, “trial and error,” or, more elegantly, direct design. We have come far with this method, cleverly combined with the instincts of skilled researchers and innovators… but it appears that we can do better. Perhaps not unlike my weekly avocado adventures.

Prof. Dequan Xiao and his research team at the University of New Haven have capitalized on our world’s rapidly increasing computational prowess and flipped the brute force process upside-down. Instead of direct design, fidgeting across an endless plane of ideas, they applied inverse design techniques to their research. For molecules, this means picking a property (such as color for avocado hunting), and backtracking to find the structure related to it (e.g. an ideal avocado) – a feasible, albeit tricky, task thanks to the relationship between chemical properties and structure.

Prof. Xiao’s work forays into vital present day problems, such as improved catalyst design for biomass to (cleaner) fuel conversion, that pave the way for easier production, lower cost, and minimized waste. Many researchers are working on this same problem from different angles, especially in “green” chemistry fields, and are making progress every day – so what makes Prof. Xiao’s approach so beneficial?

To better understand the scope of the methodology that Prof. Dequan Xiao’s team has embarked upon, imagine gazing upwards on a starry night – over 1 billion trillion (10^21) stars are estimated to be gazing back. Such an immense number is difficult to comprehend – Earth’s population hasn’t even broken 8 billion (~10^10).

What about the number of molecular design options available across the space of medicine, fuels, or materials applications?

Between 10^60 and 10^200.*

Even on the low end, this space is over 40 times larger than the number of stars in our universe! Guessing and checking our way through that many possibilities begins to look rather grim.

Prof. Dequan Xiao’s team cuts into this vast space by skipping much of the unfavorable possibilities, speeding up discovery time now that computational techniques can begin handling the load of back-calculating structure from property (albeit there are many challenges presented by this as well!).

I followed up with Prof. Xiao after his seminar series at Brown University’s chemistry department with a few broader impact questions:

1) What future impact(s) do you imagine inverse molecular design having on the drug design sector? On the materials research sector?

In the future, new computational algorithms based on Inverse molecular design theory will significantly increase the efficiency and effectiveness in searching for optimum candidate drug molecules or functional materials. Currently, the inverse molecular design theory has already shown the promise on finding candidate drug molecules, optimal optical materials, and green catalysts.  We expect that the R&D cycle of drug molecules, catalysts, or functional materials in industry will be significantly shortened due to the development of the inverse molecular design theory. More novel drugs or functional materials will be discovered to meet the needs in the market.

2) What are some of the challenges brought on that prevent inverse molecular design from being more widely implemented? (in research, industry, or otherwise?)

Designing optimal molecules or materials that can be readily accessible by experiments could be one challenge in real implementation of the inverse molecular design theory. We will need close collaboration between theoretical and experimental researchers to achieve successful chemical designs.

3) How do you see your current projects extending into the public eye? Or, what are some challenges to making new catalysts “industry ready?”

To meet the biomass conversion needs in industry, we first need to make new catalysts ‘greener’, i.e., lowering the cost while increasing the efficiency of the chemical process. The biomass catalyst that we designed is sufficiently ‘green’ (sustainable and robust) for industry applications. As the next step, we will optimize the engineering of the new catalytic process.

Inverse design theory’s promise seems clear, expediting discovery in fields ranging from catalytic materials to pharmaceuticals, but it is not without its own challenges. Fairly common across theory and design fields is the need for “close collaboration between theoretical and experimental researchers,” in order to help extricate these optimal designs from the realm of possibility and place them in the real world. Additionally, awareness of each design’s sustainability comes into play, and building in ways of ensuring “sufficiently ‘green’ (sustainable and robust)” aspects is another building block for inverse design to assimilate.

~~*~~

Prof. Dequan Xiao References of Note:

Dequan Xiao, Rui Hu, “A Tutorial of the Inverse Molecular Design Theory in Tight-Binding Frameworks and Its Applications”, Chapter 8, in Handbook of Green Chemsitry V10 – Tools for Green Chemistry, 2017.

Dequan Xiao, Ingolf Warnke, Jason Bedford, and Victor S. Batista, “Inverse Molecular Design for Materials Discovery”, RSC Specialist Periodical Report Chemical Modelling, 10, 2014, 1-31.

Dequan Xiao, Weitao Yang, and David N. Beratan, “Inverse Molecualr Design in A Tight-Binding Framework”, Journal of Chemical Physics, 129(4), 2008, 044106.

For other works by Dequan Xiao’s team, check out their publications page.

 

~~~~~~~~~~

 

*Fun fact: 10^100 is a Googol – look familiar?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s