Deciphering gp120 sequence variation and structural dynamics in HIV neutralization phenotype by molecular dynamics simulations and graph machine learning.

Human immunodeficiency virus (HIV) exploits the sequence variation and structural dynamics of the envelope glycoprotein gp120 to evade the immune attack of neutralization antibodies, contributing to various HIV neutralization phenotypes. Although the HIV neutralization phenotype has been experimentally characterized, the roles of rapid sequence variability and significant structural dynamics of gp120 are not well understood. Here, 45 prefusion gp120 from different HIV strains belong to three tiers of sensitive, moderate, and resistant neutralization phenotype are structurally modeled by homology modeling and then investigated by molecular dynamics (MD) simulations and graph machine learning (ML). Our results show that the structural deviations, population distribution, and conformational flexibility of gp120 are related to the HIV neutralization phenotype. Per-residue dynamics indicate the local regions especially in the second structural elements with high-flexibility, may be responsible for the HIV neutralization phenotype. Moreover, a graph ML model with the attention mechanism was trained to explore inherent representation related to the classification of the HIV neutralization phenotype, further distinguishing the strong related gp120 sequence variation together with structural dynamics in the HIV neutralization phenotype. Our study not only deciphers gp120 sequence variation and structural dynamics in the HIV neutralization phenotype but also explores complex relationships between the sequence, structure, and dynamics of protein by combining MD simulations and ML.

An automatic method for removing empty camera trap images using ensemble learning.

Camera traps often produce massive images, and empty images that do not contain animals are usually overwhelming. Deep learning is a machine-learning algorithm and widely used to identify empty camera trap images automatically. Existing methods with high accuracy are based on millions of training samples (images) and require a lot of time and personnel costs to label the training samples manually. Reducing the number of training samples can save the cost of manually labeling images. However, the deep learning models based on a small dataset produce a large omission error of animal images that many animal images tend to be identified as empty images, which may lead to loss of the opportunities of discovering and observing species. Therefore, it is still a challenge to build the DCNN model with small errors on a small dataset. Using deep convolutional neural networks and a small-size dataset, we proposed an ensemble learning approach based on conservative strategies to identify and remove empty images automatically. Furthermore, we proposed three automatic identifying schemes of empty images for users who accept different omission errors of animal images. Our experimental results showed that these three schemes automatically identified and removed 50.78%, 58.48%, and 77.51% of the empty images in the dataset when the omission errors were 0.70%, 1.13%, and 2.54%, respectively. The analysis showed that using our scheme to automatically identify empty images did not omit species information. It only slightly changed the frequency of species occurrence. When only a small dataset was available, our approach provided an alternative to users to automatically identify and remove empty images, which can significantly reduce the time and personnel costs required to manually remove empty images. The cost savings were comparable to the percentage of empty images removed by models.