Institute for Future Intelligence (IFI)

The material world is made of atoms and molecules. Hence, the study of molecules, or molecular sciences, is fundamentally important. The application of molecular sciences has been driving the progress of almost every field in natural science. In the past few decades, data and computation have become a powerful engine (known as in silico) for advancing molecular sciences. AIMS is IFI's innovative Web-based software solution that aims at the strategic intersection between artificial intelligence and molecular sciences or, in layman's terms, to connect bits and atoms.

AIMS currently supports two types of projects, introduced as follows. In both types of projects, AI can be used to generate novel molecules and materials that can be investigated further using molecular modeling and simulation.

Molecular Modeling

Interactive molecular dynamics allows users to design and perform computational experiments to test their own hypotheses. Play with a molecular dynamics simulation below to experience the power of computation in the browser!

Live model above (view in full screen) — Chrome or Edge recommended

With the visual analytics supported by techniques such as brushing and linking built in AIMS, users can explore quantitative structure–activity/property relationships (QSAR/QSPR) of molecules that they have collected. A well-known QSPR is the relationship between the boiling points and the carbon atom counts of alkanes (see this case study based on AIMS).

Drug Discovery

Small molecules represent an important type of drugs — 90% of all drugs sold on the global market are small molecules. The following example shows the process of screening an inhibitor to an HIV protease from a collection of small molecules using AIMS. Interactive molecular docking, currently under development, will allow users to validate a candidate molecule selected based on data science principles. For example, Lipinski's Rule of Five is implemented in AIMS as filters in the hyperspace of molecular properties as shown below.

Live model above (view in full screen) — Chrome or Edge recommended