Machine for protein binding analysis | Information dissemination media for research and development TEGAKARI

A customer involved in microbial research inquired about a machine for analyzing the binding of proteins and low-molecular-weight compounds using academic-free tools.We are planning to use it for structure prediction, docking simulation, quantum calculation, and MD calculation, and we are considering research using the following software.

Structure prediction	Alphafold2
docking	autodock (GNINA) / ucsf-dock
quantum computing	GAMESS / firefly
MD	GROMACS 2022.4

As for the specs of the desired machine, it is a condition that AlphaFold2 can be built locally and GROMACS can be operated under general conditions.
The main application is MD calculation and quantum calculation, and in the future you are thinking about developing it for in silico screening.

Based on the contents of the consultation, we proposed the following configuration.

【Main Specifications】

CPU	Xeon Gold 6326 (2.90GHz 16 cores) x2
memory	256GB REG ECC
Storage 1	2TB SSD S-ATA
Storage 2	4TB M.2 SSD
Video	Geforce NVIDIA RTX4080 x2
network	On Board (1000Base-T X2)
Housing + power supply	Tower type housing + 1600W
OS	None

It is a 3 CPU configuration equipped with 2rd generation Xeon Scalable.
This is a selection considering that some software you are using recommends the Intel API as a compiler.

Of the software you mentioned, except for GAMESS, it is possible to use the GPU in some way, but there are also some parts that are processed by the CPU, so it is necessary to consider the balance between the CPU and the GPU.

AlphaFold2 is software that requires a GPU and refers to a huge database of about 2.2 TB, so storage performance greatly affects performance. The large-capacity SSD installed as the 2nd storage is intended for AlphaFold2 database storage.

AlphaFold2 and other software use GPGPU differently. AlphaFold2 is Deep Learning based on Tensorflow, but other software is scientific calculation with GPU as the core processor.
In scientific calculations, GPGPU may be required to perform double-precision point arithmetic (fp64), but it is very expensive and requires a dedicated chassis.Since fp64 is not required for any of the software that is assumed to be used, and it is possible to use NVIDIA GPUs that support CUDA, we have selected a video card that also serves as a display output for the installed GPU.

The configuration of this case study is based on the conditions given by the customer.
Please feel free to contact us even if you are considering different conditions from what is posted.

■FAQ

・What is AlphaFold2?
AlphaFold2 is a program that predicts the three-dimensional structure of proteins by Deep Learning.Determining protein structures using conventional methods is extremely time-consuming and costly, but AlphaFold2 can perform the prediction process quickly and efficiently.Structural analysis of proteins is a valuable source of information for disease onset mechanisms and drug development.

reference:deepmind / Alphafold *Jumps to an external site

・What is GROMACS?
GROMACS (Groningen Machine for Chemical Simulations) is a molecular dynamics simulation package developed at the University of Groningen in the Netherlands.It is primarily designed for biochemical molecules such as proteins, lipids and diffusion.

reference:GROMACS *Jumps to an external site

・What is GNINA?
GNINA is a program used to estimate interactions and functional relationships between different proteins.Functional similarity between proteins can be inferred by comparing protein interaction networks.

reference:gnina / gnina *Jumps to an external site

・What is UCSF Dock?
UCSF Dock is an open source docking program for predicting interactions between proteins and small molecule ligands.It receives the molecular structural information of the ligand and protein as input, analyzes the three-dimensional structures of both, and calculates the optimal docking structure. UCSF Dock has many different functions and can be applied to numerous docking tasks such as calculation of protein-ligand interaction energies, molecular dynamics simulations, flexible docking, and virtual screening.

reference:UCSF Dock (The Official UCSF DOCK Web-site) *Jumps to an external site

・What is GAMESS?
GAMESS (General Atomic and Molecular Electronic Structure System) is computational chemistry software for electronic structural analysis of molecular structures and reactions. GAMESS can calculate physicochemical properties such as molecular electronic structures, reaction energies, and spectra using theories based on quantum mechanics.

reference:GAMES *Jumps to an external site

・What is Firefly?
Firefly (formerly known as PC GAMESS) is a computational chemistry software for electronic structural analysis of molecular structures and reactions.It supports various calculation methods, including HF, DFT, MP2, CCSD, CASPT2, and many higher-order correlation methods.These methods are used to analyze the electronic states of molecules and can calculate physicochemical properties such as reaction energies and spectra.

reference:Firefly computational chemistry program *Jumps to an external site

・What is in silico screening?
In silico screening is a method of selecting promising compounds from a large number of compounds using a computer.It is possible to input molecular structural information into a computer, conduct simulations of compound interactions and reactions, and select promising candidates from a large number of compounds.It is applied not only to drug development, but also to product development such as agricultural chemicals and cosmetics.