DESC:FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
The Patent Rules, 2003
COMPLETE SPECIFICATION
(See sections 10 & rule 13)
1. TITLE OF THE INVENTION
A MACHINE LEARNING BASED SYSTEM AND METHOD FOR CANCER TARGETED DRUG DESIGN
2. APPLICANT (S)
NAME NATIONALITY ADDRESS
DIVYASAMPARK IHUB ROORKEE FOR DEVICES MATERIALS AND TECHNOLOGY FOUNDATION IN Indian Institute of Technology Roorkee, Roorkee-247667, Uttarakhand, India.
3. PREAMBLE TO THE DESCRIPTION
COMPLETE SPECIFICATION

The following specification particularly describes the invention and the manner in which it is to be performed.

FIELD OF INVENTION:
[001] The present invention relates to the field of machine learning systems and methods for cancer targeted drug design. The present invention in particular relates to a system and method to integrate molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening for the purpose of cancer-targeted drug analysis.
DESCRIPTION OF THE RELATED ART:
[002] In the drug development process, various factors must be considered before a potential drug candidate reaches the final testing stage. Some of these factors are technical such as the rate of absorption of the drug, the duration of bioavailabilty, the administration route, its potential for side effects etc. In addition, various economic factors are also considered such as the speed and cost of the drug development process, the size of the potential market etc.
[003] The five main technical factors considered in drug development comprise:
1) Absorption of the drug (for example from the gastrointestinal (GI) tract).
2) Distribution of the drug through the body after administration (i.e. concentration in the blood stream, amount of uptake in tissues etc.).
3) Metabolism of the drug (the rate of metabolism of the drug in organs such as the liver; the metabolic stability of the drug in the body).
4) Excretion of the drug (the rate of excretion of the drug through urine or fecal matter).
5) Toxicology of the drug (what, if any, toxic side effects are exhibited by the drug).
[004] Reference may be made to the following:
[005] Publication No. US7962364B2relates to a method for performing MD simulations in drug discovery but lacks integration with dimension reduction or virtual screening techniques. The simulations are focused on protein-ligand interactions but require significant manual input and do not offer a GUI for automation or visualization like your system.
[006] Publication No. US5424963A relates to the computer-assisted method for generating a dynamic model of a molecule is described based on information of the atomic structure. The model data is defined by rigid bodies corresponding to groups of atoms of the molecule with substantially no relative movement between the atoms, flexible bodies corresponding to groups of atoms which are characterized by relative movement between the atoms, and flexure elements which define an interconnection of two of the rigid bodies and the flexible bodies and predetermined degrees of freedom.
[007] Publication No US20140074797A1 relates to a drug screening system that uses virtual screening techniques for evaluating potential drug candidates. However, it does not include molecular dynamics simulations or automated dimension reduction, which are crucial aspects of your invention's novelty.
[008] Publication No WO2015196654A1 focuses on high-throughput virtual screening but does not address MD simulations or integrate dimension reduction techniques. It primarily uses pre-docked structures without the detailed dynamic analysis provided by molecular dynamics simulations.
[009] Publication No. WO2018081354 relates to a platform and supported graphical user interface (GUI) decision-making tools for use by medical practitioners and/or their patients, e.g., to aide in the process of making decisions about a course of cancer treatment and/or to track treatment and/or the progress of a disease.
[010] Publication No. IN7812/DELNP/2014 relates to methods of "reprogramming" epigenetic mark readers or erasers to recognize epigenetic marks other than their cognate (or natural) marks are provided. Reprogramming the reader or eraser can offset the effects of aberrant writer activity (for example loss of function or overactivity) that can contribute to certain diseases states such as cancer. The use of the reprogramming compounds identified by these methods in the treatment of such disease states is also provided.
[011] Publication No. DE102004027710 relates to a method is proposed for automatically detecting a structure in an image. The method includes providing a starting region in a displayed image and prescribing a target structure. To make such a method more effective and more reliable, the concept proposed here restricts the starting region to a search region. Only then is there provision for a structure which is similar to the target structure to be automatically sought in the search region.
[012] Publication No. IN202331064842 relates to a system and method for artificial intelligence and nano technology for precision cancer medicine. The present system comprises of an artificial intelligence (AI) module designed to analyze patient-specific genomic information, pinpoint genetic mutations, and ascertain the molecular instigators of cancer.
[013] Publication No. IN3314/CHENP/2010 relates to systems, methods, and apparatus for predicting clinical outcomes and monitoring an individual’s response to a therapeutic regimen. The invention further encompasses methods for predicting cardiovascular risk based a genetic marker status and methods for modifying a computer to reflect genetic data and for incorporating genetic markers into a virtual population.
[014] Publication No. EP3347841 relates to an informatics platform provides an architecture to integrate information from relevant patient information systems. The informatics platform may include: a workflow tool that can be used to prepare and review information at multi-disciplinary board meetings; a visual timeline of patient events; a search engine to search for patients with specific attributes; a graphing tool that can display disparate clinical variables in a single chart; a virtual Pin Board for users to identify relevant patient information for board meetings; an image viewing application that can provide for comparison of images from different information systems; structured reporting functionality that incorporates system aggregated patient information and board recommendations; an application interface that integrates clinically relevant tools to provide patient specific references; a collaboration interface that facilitates communication of patient specific information and documents the discussion threads as independent reference points; and a default display of relevant patient information customized for each clinical specialty.
[015] Publication No. IN202411026601 relates to a 3D printable composite for drug delivery, process and product thereof. The composite includes a combination of 3D printable polymer including Poly-Lactic acid (PLA), a predefined drug including niclosamide, and graphene nano-platelets to provide controlled drug release through electrical stimulation. The integration of PLA as a biocompatible and sustainable polymer, along with the incorporation of niclosamide, offers a personalized and precise drug delivery system. The fabrication process involves the rotation-assisted solvent evaporation (RoSE) method, ensuring uniform drug distribution and controlled release.
[016] Patent No. US7450747 relates to a system and method for efficiently customizing an imaging system, such as a medical diagnostic imaging system. Based on images typically produced by one or more imaging systems, the present technique forms a plurality of image subject matter groups each having a plurality of related image types/views and one representative image type/view. Each customer is then provided with a plurality of predefined image style options for each of the plurality of representative image types/views.
[017] Publication No. EP1607898 relates to develops models of functional proteomics. Simulation scenarios of protein pathway vectors and protein-protein interactions are modelled from limited information in protein databases. The system focuses on three integrated subsystems, including a system to model protein-protein interactions using an evolvable Global Proteomic Model (GPM) of functional proteomics to ascertain healthy pathway operations, a system to identify haplotypes customized for specific pathology using dysfunctional protein pathway simulations of the function of combinations of single nucleotide polymorphisms (SNPs) so as to ascertain pathology mutation sources and a pharmacoproteomic modelling system to develop, test and refine proposed drug solutions based on the molecular structure and topology of mutant protein(s) in order to manage individual pathologies.
[018] Patent No. US6970791 relates to a method of carrying out a molecular modeling calculation and displaying the results thereof through a task-oriented user interface. The overall computational task is divided into subtasks, each of which is outfitted with a tailored graphical user interface. In a preferred implementation, the subtask user interfaces are accessed via tab icons whose layout reflects the normal order of carrying out the subtasks, and downstream subtask user-interfaces are not activated until the information they require is available from upstream subtasks.
[019] Publication No. US2015269764 relates to systems that allow users to design and model natural structures such as biomolecules. The system can include a collection of individual models and also provide users with the option to select certain simulation/interaction modalities that will influence the dynamics of models within a simulation created by the user.
[020] Patent No. US11237713 relates to a mechanism is provided in a data processing system to implement a feature extraction tool for graphical user interface based feature extraction. The feature extraction tool receives selection by a user of a dataset from which features are to be extracted. The feature extraction tool loads a plurality of feature definitions. The feature extraction tool generates a graphical user interface that allows the user to add features from the plurality of features to a feature file.
[021] Patent No. US8880350 relates to a system and method for determining individualized medical intervention for a particular disease state, and especially for cancers, that includes the molecular profiling of a biological sample from the patient, determining whether any molecular findings including one or more genes, one or more gene expressed proteins, one or more molecular mechanisms, and/or combinations of such exhibit a change in expression compared to a reference, and identifying a non-specific disease therapy or agent capable of interacting with the genes, gene expressed proteins, molecular mechanisms, or combinations of such molecular findings that exhibited a change in expression.
[022] Publication No. JP2007334801 relates to a patient information integrated drawing system in which the content of medical treatment received by a patient can be sufficiently clearly explained. The patient information integrated drawing system has a plurality of servers including a medical image server, a hospital information server and a drug information server, an information terminal which is connected to the plurality of servers and gives instructions to collect information about a specific patient and a patient information integrated drawing device which stores the information about the specific patient which the information terminal instructs to collect and displays the information about the specific patient stored in the patient information integrated drawing device on an image display means via the information terminal.
[023] Patent No. US10636516 relates to systems, methods, and computer-readable media for analyzing and presenting healthcare information are described. Some embodiments may include a system configured to receive healthcare information relating to a patient and to generate a patient profile. The patient profile may include a physiological status as well as a physiological assessment and a treatment assessment based on the automatic and dynamic analysis of the healthcare information.
[024] Publication No. IN4907/DELNP/2007 relates to a computer-implemented method for calculating a value representative of interaction (VRI) of a proposed ligand with a specified receptor. Hydrophobic interactions between one or more ligand atoms and one or more receptor atoms are scored by a method that awards a bonus for the presence of hydrophobic enclosure of one or more ligand atoms by the receptor.
[025] Publication No. IN202441024996 relates to our research focuses on optimizing structures to selectively target HER2 kinase, a crucial target in breast cancer therapy linked to aggressive clinical outcomes.
[026] Publication No. IN202411082441 relates to a novel Insilico methodology for identifying potential inhibitors targeting key molecular drivers in retinoblastoma, a highly malignant pediatric eye cancer. Retinoblastoma cells exhibit aberrant activity of Polo-like kinase 1 (PLK1) and Cyclin-dependent kinase 4 (CDK4), which are crucial for tumour progression and cell cycle regulation. This approach utilizes advanced computational techniques to screen and evaluate the efficacy of compounds 2760 and 1950 for PLK1 inhibition and compounds 3396 and 960 for CDK4 inhibition.
[027] Publication No. WO2023007360 relates to a molecule for targeting the specific receptors in mitigating breast cancer and prostate cancer. The magic bullet is derived as a small molecule using the molecular modelling technique. The molecule of formula (I) comprises the acid as first chemical moiety, hormone derivative as second chemical moiety and the anticancer agent as third chemical moiety.
[028] Publication No. EP2618102 relates to a system, apparatus and method of obtaining data from a 2D (two-dimensional) image in order to determine the 3D (three-dimensional) shape of objects appearing in said 2D image, said 2D image having distinguishable epipolar lines, said method comprising: providing a predefined set of types of features, giving rise to feature types, each feature type being distinguishable according to a unique bi-dimensional formation; providing a coded light pattern comprising multiple appearances of said feature types; projecting said coded light pattern on said objects such that the distance between epipolar lines associated with substantially identical features is less than the distance between corresponding locations of two neighboring features; capturing a 2D image of said objects having said projected coded light pattern projected thereupon, said 2D image comprising reflected said feature types; and extracting: said reflected feature types according to the unique bi-dimensional formations; and locations of said reflected feature types on respective said epipolar lines in said 2D image
[029] Traditional MD simulations require specialized knowledge, manual configuration, and extensive computational resources.
[030] Principal Component Analysis (PCA) and other reduction techniques are vital for simplifying high-dimensional MD data but traditionally require expertise in statistics and manual coding.
[031] Existing virtual screening processes are fragmented, often requiring multiple tools and manual file format conversions, which can lead to inefficiencies.
[032] In order to overcome above listed prior art, the present invention aims to provide a machine learning based system and method which integrates molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening for the purpose of cancer-targeted drug discovery. The invention provides a user-friendly GUI that automates complex setups, making it accessible to non-experts. It supports multiple force fields and allows for the adjustment of simulation parameters, improving usability and reducing time-intensive setup steps. The invention automating the process and allowing visualization of reduced data in 2D/3D plots for easier interpretation of key molecular features.
[033] This automated process will operate without requiring prior knowledge of these techniques. The invention integrates a virtual screening module that automates the evaluation of large compound libraries and supports scoring functions like docking scores and physicochemical filters, streamlining the process.
OBJECTS OF THE INVENTION:
[034] The principal object of the present invention is to provide a system and method to integrate molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening for the purpose of cancer-targeted drug discovery.
[035] Another object of the present invention is to provide machine learning based systems and methods for cancer targeted drug design.
[036] Yet another object of the present invention is to provide a system which reduces time and cost involved in discovering new drug molecules targeting selected kinases.
[037] Still another object of the present invention is to provide an integrated system that improves efficiency, reduces the technical barrier, and accelerates drug discovery through automation and streamlined processes.
SUMMARY OF THE INVENTION:
[038] The present invention relates to the machine learning based systems and methods for cancer targeted drug design. The system integrates molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening for the purpose of cancer-targeted drug. The system automates complex computational workflows, providing protein-ligand interactions, reduce high-dimensional molecular data, and perform virtual screening to identify potential drug candidates.
[039] The system includes the invention is a machine learning-based system that integrates Molecular Dynamics (MD) simulations, dimension reduction techniques, and virtual screening into a single user-friendly Graphical User Interface (GUI). It automates traditionally labor-intensive workflows, making drug more accessible and efficient.
BREIF DESCRIPTION OF THE INVENTION
[040] It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered for limiting of its scope, for the invention may admit to other equally effective embodiments.
[041] Figure 1 shows systems for cancer targeted drug design.
[042] Figure 2 shows flowchart according to the present invention.
DETAILED DESCRIPTION OF THE INVENTION:
[043] The present invention provides the machine learning based systems and methods for cancer targeted drug design. The system integrates molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening for the purpose of cancer-targeted drug. The system automates complex computational workflows, providing protein-ligand interactions, reduce high-dimensional molecular data, and perform virtual screening to identify potential drug candidates.
[044] The system includes the invention is a machine learning-based system that integrates Molecular Dynamics (MD) simulations, dimension reduction techniques, and virtual screening into a single user-friendly Graphical User Interface (GUI). It automates traditionally labor-intensive workflows, making drug more accessible and efficient.
[045] Referring to figure 1, system comprises control unit (1), input unit (2), display unit (3), memory (4), and communication interface (5). Control Unit (1) coordinates all operations, processes inputs, and manages the flow between MD simulations, PCA, and virtual screening. Input unit (2) allows users to upload files (PDB, ligand libraries, etc.), define parameters (temperature, pressure), and select tasks (e.g., MD run, PCA, docking). Display unit (3) Graphical interface that shows MD simulation visuals, PCA plots, docking scores, and status updates. Memory (4) stores input/output files, intermediate simulation data, and processed results for user review. Communication Interface (5) facilitates integration with external software like GROMACS or AutoDock, and potentially cloud or hardware accelerators.
[046] Input Unit (2) feeds data to the Control Unit (1), which initiates simulations, reduction, or docking. Control Unit (1) processes the instructions and interacts with external simulation tools through the communication interface (5). intermediate and final results are stored in memory (4) and visualized through the display unit (3). The virtual screening module operates either concurrently or post-PCA, using structures from MD simulations to screen ligands. All components work in sync under the orchestration of the Control Unit, and are accessible through a user-friendly GUI.
[047] The invention is a user-friendly GUI based system that automates complex setups, making it accessible to non-experts. It supports multiple force fields and allows for the adjustment of simulation parameters, improving usability and reducing time-intensive setup steps. The system incorporates Principal Component Analysis (PCA) and other reduction techniques into the GUI, automating the process and allowing visualization of reduced data in 2D/3D plots for easier interpretation of key molecular features.
[048] Molecular dynamics (MD) simulation module allows to set up and run MD simulations to analyze protein-ligand interactions. It supports tools like GROMACS with integrated force fields such as AMBER and CHARMM. The users can adjust parameters (e.g., temperature, pressure) via the GUI without manual coding and provides real-time feedback and visualization of simulations to ensure dynamic adjustments during runtime.
[049] It automates the simplification of high-dimensional data using techniques like Principal Component Analysis (PCA), visualizes reduced data in 2D/3D plots, highlighting conformational changes and trends in molecular structures and enables non-experts to identify key molecular features without statistical expertise. The PCA module reduces complex, high-dimensional MD simulation data into fewer dimensions (2D or 3D) that retain most variance. This reveals essential molecular motions (e.g., domain opening/closing), aiding in visual interpretation of protein behavior during simulations. Non-experts can identify biologically significant patterns or conformational clusters without writing code.
[050] Thus the screens large libraries of compounds to evaluate binding affinities and drug-likeness and integrates docking software to automate scoring and filtering processes. It facilitates efficient identification of potential drug candidates by reducing manual file conversions and multiple-tool dependencies. This will significantly reduce the time and cost involved in new drug molecules targeting selected kinases. This automated process will operate without requiring prior knowledge of these techniques. The invention integrates a virtual screening module that automates the evaluation of large compound libraries and supports scoring functions like docking scores and physicochemical filters, streamlining the process. The virtual screening module screens thousands of compounds for binding with the target protein. It automates:
• Docking each compound
• Evaluating docking scores (binding affinity)
• Applying filters (e.g., Lipinski’s Rule of 5)
[051] This eliminates manual work and enables fast, large-scale screening to find promising leads for further testing.
[052] The system includes integration of molecular dynamics (MD) simulations, dimension reduction techniques, and virtual screening into a single, user-friendly Graphical User Interface (GUI). The GUI connects modules (MD simulation, PCA, and virtual screening) in a single interface. Users can:
1. Upload inputs
2. Run simulations
3. Reduce data using PCA
4. Perform virtual screening
5. View plots and results
[053] This integrated approach eliminates switching between tools, reduces file conversion needs, and accelerates analysis.
[054] This integration streamlines what are traditionally separate, labor-intensive tasks, making the complex drug process more accessible, efficient, and automated.
[055] The system allows non-computational biologists and researchers without programming skills to perform advanced computational tasks. In most related fields, it would be expected that MD simulations and drug screening require a deep understanding of computational biology or scripting languages. Offering this functionality in an intuitive interface is an inventive step that addresses accessibility issues without compromising on the system’s technical robustness.
[056] Providing real-time feedback from MD simulations, dimension reduction, and virtual screening allows researchers to dynamically adjust their analysis and workflows. This real-time integration of separate processes is non-obvious to those in related fields because typical systems require post-simulation manual data processing, where real-time analysis is often not feasible due to the complex and high-dimensional nature of the data involved.
[057] The specific focus on cancer-targeted drug development, with built-in support for cancer-specific protein databases and molecular targets. The system integrates MD simulations, dimension reduction, and virtual screening into a unified GUI. This seamless integration eliminates the need for multiple software tools, reducing errors and inefficiencies. The system provides on-screen updates during simulations and screening. For example:
• RMSD graphs update live as MD progresses
• PCA plots are generated immediately after selection
• Docking scores appear upon completion
[058] This helps users adjust parameters or rerun simulations without delay, a key time-saving feature.
[059] It automates complex computational processes, making them accessible to researchers without prior programming or statistical expertise. The system is preloaded with cancer-relevant targets (e.g., CDK2, EGFR, HER2) and their structures. It allows users to select proteins implicated in cancers, ensuring relevance. This focus on cancer proteins provides a tailored drug discovery pipeline compared to general-purpose tools.
[060] It enables real-time feedback and dynamic adjustments, which are non-obvious features in traditional systems that require post-simulation manual processing.
[061] This is tailored for cancer-targeted drug development, with support for cancer-specific protein databases and molecular targets. It offers enhanced relevance to a critical area of drug research, which is not addressed in general-purpose tools.
[062] Thus present invention makes advanced computational techniques portable and cost-effective. It provides a balance between computational power and affordability, addressing the limitations of traditional high-end server-based systems.
[063] It simplifies the setup of simulations, data reduction, and screening processes with an intuitive interface. This GUI-based approach reduces the technical barrier for researchers and experimental biologists, enabling faster adoption.
[064] The automated system provides real-time integration, and hardware optimization ensuring that it addresses key challenges in drug development, including accessibility, cost, and efficiency. It represents a significant advancement over traditional methods, providing a powerful tool for researchers in cancer-targeted drug development.
[065] The invention integrates Molecular Dynamics (MD) simulations, Principal Component Analysis (PCA), and virtual screening into a single, user-friendly graphical interface. This eliminates the need for multiple software tools, file conversions, and complex command-line interactions typically required in prior art.
[066] The system provides real-time RMSD updates, PCA visualizations, and docking results during the simulation process, allowing users to make on-the-fly adjustments. This dynamic, live feedback mechanism is a significant improvement over traditional post-simulation data processing approaches.
[067] It is designed with no coding requirement, the GUI makes advanced computational techniques accessible to wet-lab researchers, clinicians, and life scientists without a background in computational biology or scripting.
[068] The system includes built-in support for cancer-specific protein targets (e.g., kinases like CDK2, EGFR), making it specialized for oncology-focused drug, which is not addressed in general-purpose platforms.
[069] A notable inventive feature is the system’s ability to run efficiently on edge hardware, such as the NVIDIA Jetson Orin Nano Developer Kit. This enables portable, low-cost deployment, unlike prior systems which require high-end server infrastructures. It broadens accessibility to institutions with limited computational resources, including remote labs and field research units.
[070] The following ZINC compounds were identified after molecular docking based on their strong binding energies (MolDock Score). Their IUPAC names are provided below:
1. ZINC12369719 – 5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole
2. ZINC04394657 – N-[(2-chlorophenyl)methylideneamino]-4-(5-methyl-3-phenyl-1,2-oxazol-4-yl)-1,3-thiazol-2-amine
3. ZINC01047457 – 2-(3,4-dichlorophenyl)-N-(thiophene-2-carbonyl)-1,3-thiazole-4-carbohydrazide
4. ZINC00169253 – 4-[2-(2,1,3-benzoxadiazol-5-yloxy methyl)-1,3-thiazol-4-yl]benzonitrile
5. ZINC13658326– 4-[(Z)-(3-(4-chlorophenyl)-1-phenylpyrazol-4-yl)methylideneamino]-3-thiophen-2-yl-1H-1,2,4-triazole-5-thione
6. ZINC12369719 (duplicate) – 5-[3-methyl-5-(4-methylthiadiazol-5-yl)-1,2-oxazol-4-yl]-3-(5-nitrothiophen-2-yl)-1,2,4-oxadiazole
[071] The method includes following
1) Input Step:
• User uploads protein structure (PDB)
• Selects ligand database and simulation parameters via GUI
2) MD Simulation:
• Backend uses GROMACS with selected force field
• Performs equilibration, production runs
• Trajectories are saved
3) Data Reduction:
• PCA is applied to trajectories
• Reduced data plotted in 2D/3D
• Key motion patterns visualized
4) Virtual Screening:
• Ligands are docked to selected protein conformers
• Docking scores calculated automatically
• Filters applied (e.g., Lipinski, ADMET)
5) Output & Decision Support:
• GUI displays ranked ligands
• Results downloadable for further testing
[072] The invention integrates all these modules into a cohesive graphical interface. This seamless integration simplifies the overall workflow, reducing errors, saving time, and increasing usability—particularly for non-computational users.
[073] The invention is specifically designed for cancer-targeted drug discovery, incorporating support for oncology-relevant targets like CDK2, EGFR, HER2, etc., and built-in access to cancer-specific protein databases. Existing tools are often general-purpose and not optimized for oncology use cases.
[074] The system is deployable on cost-effective platforms such as the NVIDIA Jetson Orin Nano Developer Kit, allowing advanced computations to be run on compact, energy-efficient devices. This enhances portability, enabling use in remote labs or educational institutions with limited infrastructure.
[075] Numerous modifications and adaptations of the system of the present invention will be apparent to those skilled in the art, and thus it is intended by the appended claims to cover all such modifications and adaptations which fall within the true spirit and scope of this invention.
,CLAIMS:WE CLAIM:
1. A machine learning based system for cancer targeted drug design comprises-
a) control unit (1), processes the instructions and interacts with external simulation tools through the communication interface (5),
b) input unit (2), feeds data to the control unit (1), which initiates simulations, reduction, or docking,
c) display unit (3), shows MD simulation visuals, PCA plots, docking scores, and status updates,
d) memory (4), storing intermediate and final results in memory (4) which are visualized through the Display Unit (3),
e) communication interface (5),
f) a virtual screening module that automates the evaluation of large compound libraries and supports scoring functions like docking scores and physicochemical filters, streamlining the process.
2. The machine learning based system for cancer targeted drug design, as claimed in claim 1, wherein the cancer-targeted drug development is provided with built-in support for cancer-specific protein databases and molecular targets including preloaded or configured to interface with databases that contain structural and functional information of proteins known to be involved in various forms of cancer and focuses on disease-specific targets, the system improves the biological significance of the drug candidates identified and reduces false positives that may occur when using general-purpose tools.
3. The machine learning based system for cancer targeted drug design, as claimed in claim 1, wherein the system integrates MD simulations, dimension reduction, and virtual screening into a unified GUI which automatically screens compounds, calculates binding scores, and filters hits using physicochemical properties and docking outputs are displayed in a user-friendly manner.
4. The machine learning based system for cancer targeted drug design, as claimed in claim 1, wherein the system automates the evaluation of large compound libraries and supports scoring functions like docking scores and physicochemical filters, streamlining the process.
5. The machine learning based method for cancer targeted drug design, includes following steps:
a) Input Data Collection providing Protein target structure (e.g., CDK2.pdb file), Ligand library (from sources like the ZINC database), simulation parameters (e.g., temperature, pressure, duration) and data is uploaded via the GUI’s Input Unit.
b) Molecular Dynamics (MD) Simulation.
c) The high-dimensional MD trajectory data is processes.
d) The system extracts and visualizes key molecular motions in 2D or 3D.
e) The system automatically docks the ligand library against representative protein conformations obtained from PCA or MD and each ligand is scored based on binding affinity and filtered using drug-likeness criteria.
f) Output includes a ranked list of compounds with docking scores and final results, including top-ranked ligands and protein-ligand interactions, are displayed and users can download data for further analysis or experimental validation.