Adding hydrogen atoms to molecular models via fragment superimposition

Area of Science:

• Biochemistry
• Structural Biology
• Computational Chemistry

Background:

• Experimentally determined biomolecular structures often lack hydrogen atom positions due to low electron density.
• Accurate hydrogen atom positions are crucial for detailed structure analysis and molecular simulations.
• Current hydrogen prediction methods have limitations in scope and effectiveness for diverse molecular systems.

Purpose of the Study:

• To develop a novel algorithm for predicting hydrogen atom positions in organic compounds and biomacromolecules.
• To create an accessible and efficient computational tool for structural biologists and biophysicists.

Main Methods:

• A library of molecular fragments with known hydrogen atom positions is compiled.
• The algorithm predicts hydrogen positions in new molecules by referencing similar moieties in the library.
• The method's accuracy is evaluated across various organic compounds, including biomacromolecules.

Main Results:

• The novel algorithm accurately predicts hydrogen atom positions for a wide range of organic compounds.
• The method demonstrates effectiveness for biomacromolecules when utilizing a sufficiently large fragment library.
• The developed algorithm forms the basis of the open-source Hydride software package.

Conclusions:

• The Hydride software package, implemented as an open-source Python program, provides an out-of-the-box solution for hydrogen position prediction.
• The software requires minimal to no additional parameterization, enabling rapid computation (typically seconds).
• Hydride is presented as a valuable and efficient tool for researchers in structural biology and biophysics.