RAIN: a Machine Learning-based identification for HIV-1 bNAbs.

Area of Science:

• Immunology
• Computational Biology
• Structural Biology

Background:

• Broadly neutralizing antibodies (bNAbs) are crucial for developing effective HIV-1 treatments and prevention strategies.
• Current methods for identifying HIV-1 bNAbs from immune repertoires are limited, hindering rapid discovery.
• Automatic detection of bNAbs from immune repertoire data is a significant unmet need.

Purpose of the Study:

• To develop a straightforward computational method for the rapid and automatic identification of HIV-1 bNAbs.
• To overcome limitations of existing methods by utilizing sequence-based features for prediction.
• To accelerate the discovery of novel HIV-1 bNAbs from non-selected immune repertoires.

Main Methods:

• Developed RAIN (Rapid Automatic Identification of bNAbs), a Machine Learning-based computational method.
• Utilized a combination of selected sequence-based features for predicting HIV-1 bNAbs, differing from one-hot encoding or structural alignment.
• Applied RAIN to non-biased, experimentally obtained sequenced BCR repertoires from HIV-1 immune donors.

Main Results:

• RAIN successfully identified novel HIV-1 bNAbs targeting the CD4-binding site of the HIV-1 envelope glycoprotein.
• Validated the identified bNAbs through in vitro neutralization assays.
• Determined the structure of a novel bNAb in complex with the HIV-1 envelope glycoprotein using single-particle cryo-electron microscopy (cryo-EM).

Conclusions:

• RAIN provides a facile and accelerated method for discovering HIV-1 bNAbs.
• The developed method facilitates the identification of bNAbs from non-selected immune repertoires.
• This work advances the potential for developing new HIV-1 therapeutics and preventative measures.