Chemistry Question

TheoryLeft: Protein Right: CASP14 contest (Callaway)
DeepChem predicts the reactant molecules
required to produce a product molecule, given
the type of reaction, using a neural network
trained on a dataset of millions of known
reactions. This task is known as retrosynthetic
reaction prediction (Ramsundar).
Retrosynthetic prediction (Ramsundar)
DeepMind’s AlphaFold used neural networks to
predict protein-folding much more accurately
than previous winners of the CASP14 contest,
though it faces criticism from scientists for its
lack of transparency (Callaway).
New Technology
ORF3a (Bank)
Drug discovery
using AI is an
academic field
which is growing
exponentially
quickly (Belenzon,
Alpaydin, Ethem. Introduction to machine learning. MIT press, 2020.
Bank, RCSB Protein Data. “6XDC: Cryo-EM Structure of SARS-CoV-2 ORF3a.” RCSB PDB, 2020, www.rcsb.org/structure/6XDC.
Belenzon, Liran, and Simon Smith. “ARTIFICIAL INTELLIGENCE – 3Ds Powering AI in Drug Discovery – Domain Expertise, Deep
Learning & Data.” Drug Development and Delivery, 4 Oct. 2018, drug-dev.com/artificial-intelligence-3ds-powering-ai-indrug-discovery-domain-expertise-deep-learning-data/.
Callaway, Ewen. “’It Will Change Everything’: DeepMind’s AI Makes Gigantic Leap in Solving Protein Structures.” Nature News, Nature
Publishing Group, 30 Nov. 2020, www.nature.com/articles/d41586-020-03348-4.
DeepAI. “Hidden Layer.” DeepAI, DeepAI, 17 May 2019, deepai.org/machine-learning-glossary-and-terms/hidden-layer-machinelearning.
Freedman, David H. “Hunting for New Drugs with AI.” Nature News, Nature Publishing Group, 18 Dec. 2019,
www.nature.com/articles/d41586-019-03846-0.
Paul, Debleena, et al. “Artificial Intelligence in Drug Discovery and Development.” Drug Discovery Today, Elsevier Ltd., Jan. 2021,
www.ncbi.nlm.nih.gov/pmc/articles/PMC7577280/.
West, Darrell M., and John R. Allen. “How Artificial Intelligence Is Transforming the World.” Brookings, Brookings, 28 Apr. 2020,
www.brookings.edu/research/how-artificial-intelligence-is-transforming-the-world/.
Yobero, Czar. “K-Means Clustering Tutorial.” RPubs, 2018, rpubs.com/cyobero/k-means.
Citations
Growth of AI in Drug Discovery (Belenzon) Freedman).
AlphaFold was successfully
able to approximately
predict the structure of the
SARS-CoV-2 ORF3a
transport protein before
experimental analysis using
cryo-EM was complete
(Bank).
Applications
Artificial Intelligence in Chemistry
Artificial Intelligence—A machine which has
intentionality, intelligence, and adaptability
Neural Network (DeepAI)
In supervised learning, an existing labelled data
is used to predict difficult-to-measure outputs
based on easier-to-measure inputs, for example
by training a neural network (Alpaydin).
K-means clustering (Yobero)
In unsupervised learning, unlabeled data is used
to form categories which describe different
aspects of the data, for example using k-means
clustering (Alpaydin).
Introduction [1,2]:
References:
1.
2.
3.
4.
5.
https://pubs.acs.org/doi/10.1021/acsomega.0c01950
https://www.sciencedirect.com/science/article/pii/S2405829719302922
https://pubs.acs.org/doi/pdf/10.1021/acsami.8b11824
https://www.sciencedirect.com/science/article/pii/S2542435118304069
https://doi.org/10.1002/adma.201700519
There are two main subtle differences between conventional rocking-chair batteries and DIBs:
1) Insertion mechanism
2) Charge carriers
In rocking-chair batteries, cations are responsible for electrochemical reactions at both electrodeelectrolyte interfaces however, in DIBs, the electrolyte provides charge carriers and both cations and
anions are responsible for electrochemical reactions and they are being captured from and released to
the electrolyte [4,5].
Schematic structure of conventional rocking-chair batteries vs. DIBs [4].
In DIBs, both cation and anions are involved in the redox reaction.
DIBs schematic structure is shown below. DIBs have similar configuration as conventional rocking-chair
batteries with two electrodes separated by a polymer membrane and electrolyte [4].
Dual-Ion Batteries (DIBs) Theory and Structure:
Week 1 – Assignment 2
Top Emerging Technologies in Chemistry
Dual Ion Batteries (DIBs)
Afshin Gholamy
Dual-Ion Batteries (DIBs) with different electrochemical configurations are emerging energy storage
systems. DIBs history goes back to advent of graphite intercalation compounds (GICs) in 1938. Earlier
versions of DIBs were known as dual graphite batteries (DGBs) or dual carbon batteries (DCBs) since
they used graphite for both electrodes. DIBS suffered from the following issues:
1) Electrolyte decomposition at working voltages of > 4.5V
2) Graphite exfoliation
3) Low capacity
4) Scarcity of Li for Li-based DIBs
LIBs vs. DIBs [1,2,3]:
Lithium-Ion Batteries are energy storage solution for a wide range of applications such as handheld
electronic devices however there are two major drawbacks associated with LIBs:
1)
They rely on lithium and some other scarce elements
2)
They are only applied to low capacities/low working voltage applications (3.6-3.85 v).
Unlike, Lithium-Ion Batteries (LIBs), Dual-Ion Batteries (DIBs)
1) Rely mainly on abundant elements such as Na, K, Zn and Al which makes them
environmentally friendly.
2) Have a lower cost
3) Higher lifetime, safety, and sustainability
4) Higher performance: High energy density, capacity and working voltages.
5) High-speed charge/discharge capability at large currents which makes them appealing to
a wide range of high-power applications such as grid powers, electric vehicles and heavy
machinery. This capability stems from low internal resistance of DIBs.
6) Ease of recycling
DIBs: Extrapolating into the Future [1,3,4]:
The future of energy storage is focused on improving the performance of batteries by incorporating
multiple (triple, quadruple) ions configurations. This would be beneficial to diffusion kinetics, working
voltage, cycling reversibility and more electrode options.
Scientists are also looking into improving the conventional liquid/polymer gel electrolytes and even
replacing them with solid electrolytes such as ceramics, glass, sulfites which would improve the
thermal efficiency and safety, energy density, life expectancy and lower mass production cost.
Some challenges that scientists need to address with future generation of batteries include:
1)
Sluggish kinetics
2)
Poor reversibility
3)
Lack of suitable material for electrodes.