The Human Immunodeficiency Virus, also known as HIV, is a retrovirus that causes deficiencies in the human immune system. This virus attacks and gradually destroys the human immune system, leaving the host unprotected during infection. People who are infected with HIV and pass away often die from secondary infections or cancer. AIDS is the final stage of HIV infection.
There are two main types of the AIDS virus: HIV-1 and HIV-2. HIV-1 originated in the area around the Congo Basin in Africa and is the most prevalent strain globally. It is responsible for about 95% of all HIV infections. HIV-2 is primarily found in West Africa, although it also affects a small number of people in Europe, India, and the United States.
HIV-1 can be divided into 4 groups: M, N, O and P, and Group M is the most widely distributed worldwide, which is divided into 12 different subtypes or sub-subtypes, namely A1, A2, B, C, D, F1, F2, G, H, J, K, and L. The formation of various subtypes or sub-subtypes of HIV-1 group M is the result of continuous molecular evolution. Correct classification of subtypes or sub-subtypes is important for vaccine design, therapeutic drugs, and effective prevention and control of AIDS in the global community.
Precise M subtype or sub-subtype classification relies on phylogenetic analysis of specific gene sequences. In the past, subtypes were roughly determined through homologous searches in the NCBI database. The accuracy depended on the searcher's judgment level, and sometimes even the correct judgment could not be made. Classification of HIV-1 subtype based on statistical modeling methods has also been developed, but due to the small sample size of some subtypes, this has caused great limitations in the tools derived from statistical modeling. Therefore, I developed a deep learning-based method, which I named HIV-1-M-SPBEnv.
To address the issue of samples being scarce for some subtypes, I successfully used artificial genetic mutation methods to synthesize new machine learning samples, thereby cleverly solving the problem of insufficient samples for some subtypes.
Due to the rapid evolution rate of the env gene, I chose to model the env gene sequence samples for the classification of HIV-1 subtypes. I download env gene sequences at the HIV Sequence Database supported by Los Alamos National Laboratory (https://www.hiv.lanl.gov/content/index). I downloaded the 2022 version of the env gene sequence, which is the latest version so far.
In my deep learning model framework, I use the Kmer method to vectorize DNA sequences; The model training adopts a unique strategy. The function of the Autoencoder is to extract high-dimensional feature information from DNA sequences. There are two training cycles, including Autoencoder training cycle and classifier training cycle. In the Autoencoder training cycle, I train the Autoencoder with the criterion of making the output reconstruction loss approach zero, thereby enabling the output signal to completely reconstruct the input signal, while in the classifier training cycle, the goal is to train a classifier to minimize its training loss, thereby achieving higher classification accuracy. To this end, the two loss function values are added together to obtain a total loss function, and this total loss function is made to approach zero after the completion of training; this serves as the training criterion for the entire model. With the total loss approaching zero, the Autoencoder's reconstruction loss also tends towards zero, then achieving a perfect reconstruction of the Autoencoder's output to its input. Consequently, the encoder's output thus perfectly and losslessly extracts high-dimensional feature information of DNA sequences, representing the high-dimensional information of DNA sequences with low-dimensional information. Subsequently, the output signal of the encoder is fed into the full connected neural network block, which is used for the classification task.
Using an independent validation dataset, the accuracy of HIV-1-M-SPBEnv reached 100%, demonstrating strong model generalization capabilities. The trained model is deployed at http://www.hivsubclass.com/.
Python == 3.10.13
torch == 2.1.0+cu121
torchvision == 0.16.0+cu121
scikit-learn == 1.3.2
pandas == 2.1.2
numpy == 1.24.1
matplotlib == 3.8.0
biopython == 1.79
The following commands only use the Windows system as an example.
conda create -n HIV-1-M-SPBEnv python=3.9.18
conda activate HIV-1-M-SPBEnv
# My GPU Cuda version is 12.3. However the latest version of Cuda compatible with Pytorch is Cuda 1.21.
# Please check at https://pytorch.org/ for appropriate installation commands for your computer.
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
# If you do not have a GPU card on your computer, please use the following command to install Pytorch:
pip3 install torch torchvision torchaudio
#Clone the HIV-1-M-SPBEnv repo
git clone https://github.com/pengsihua2023/HIV-1-M-SPBEnv.git
cd HIV-1-M-SPBEnv
# Install requirements
pip install -r requirements.txt
#Install HIV-1-M-SPBEnv
pip install .#Just execute the command "hiv-env" in any command line environment.
hiv-envpip uninstall HIV-1-M-SPBEnvI adopted artificial molecular evolution methods for DNA sequence sample synthesis, including synonymous mutation, non-synonymous mutation, insertion mutation, deletion mutation and large fragment recombination. The complete codebase for artificial genetic operations is accessible at the following link: https://github.com/pengsihua2023/HIV-1-M-SPBEnv/tree/main/code
I deployed a trained model on a dedicated server, which is publicly available at:
http://www.hivsubclass.com/, to make it easy for biomedical researcher to perform HIV 1 subtype classification in their research.
Users can upload their env gene sequences of HIV to the server, and then they can quickly obtain the predicted results of the HIV 1 subtype classification.
HIV-1-M-SPBEnv has a MIT license, as seen in the LICENSE.md file.
Sihua Peng, Ming Zhang. HIV-1-M-SPBEnv: HIV-1 M group subtype prediction using convolutional Autoencoder and full connected neural network. (Manuscript under review)
If you have any questions, please feel free to contact Sihua Peng (Email: Sihua.Peng@uga.edu)
Pull requests are highly welcomed!