A hybridized gravitational search based system for planning blood-brain barrier penetrating molecule
Patent Information
- Application Number
- IN202511033030
- Authority / Receiving Office
- IN · IN
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2025-04-03
- Publication Date
- 2026-10-23
- Estimated Expiration
- 2045-04-03
AI Technical Summary
Current methods for developing blood-brain barrier (BBB) penetrating molecules face challenges such as lack of specificity, systemic toxicity, limited drug efficacy due to saturable transport mechanisms, high production costs, and complexity in large-scale manufacture, which hinder accurate, safe, and scalable medicinal delivery to the brain.
A hybridized gravitational search based system that utilizes explainable artificial intelligence modules to discover and optimize peptides for BBB penetration, determining critical amino acids and optimizing characteristics through a hybridized gravitational search protocol, thereby reducing computational time and cost while addressing multiobjective optimization.
The system effectively identifies blood-brain barrier penetrating peptides with desirable characteristics, such as high isoelectric point and low molecular weight, enhancing therapeutic efficacy while minimizing off-target effects and production costs.
Abstract
Description
FIELD OF THE INVENTION The present invention pertains to a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules. More particularly, the present invention provides a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules. BACKGROUND OF THE INVENTIONThe blood-brain barrier (BBB) is a semi-permeable barrier that surrounds the central nervous system (CNS). The endothelial cells that make up the highly specialized, selectively permeable interface are joined by tight junctions, astrocytic end-feet, and a supporting basement membrane. As a vital defense mechanism, the endothelial cells protects the central nervous system (CNS) by limiting the infiltration of harmful substances, pathogens, and huge molecules while facilitating the controlled circulation of essential gases and nutrients. This barrier is essential for preserving the homeostasis of the brain and promoting the optimal possible neuronal function.Despite the protective benefits, the BBB exhibits substantial barriers to the administration of therapeutic drugs intended for the treatment of CNS diseases. To address this, researchers have identified and developed blood-brain barrier-penetrating molecules. These molecules are specifically developed or selected to transverse the BBB, leveraging various mechanisms such as passive diffusion, carrier-mediated transport (e.g., glucose transporters), or receptor-mediated transcytosis (transferrin or insulin pathways). While peptides and biologics make use of receptor-mediated routes, lipophilic substances with low molecular weight frequently display passive diffusion. The development of targeted compounds, such as prodrugs, lipid nanoparticles, and antibody-drug conjugates, to improve BBB permeability has also been made possible by developments in medicinal chemistry and nanotechnology. Furthermore, endogenous transport pathways are being investigated in targeted delivery systems to enable precise drug delivery with minimal off-target effects. Blood-brain barrier penetrating molecules face several significant limitations, impacting their therapeutic efficacy. Their lack of specificity is a significant disadvantage, as many of these substances indiscriminately invade other tissues, resulting in systemic toxicity and off-target consequences. Furthermore, the different type of therapeutic drugs that are efficiently transported to the central nervous system are limited by the BBB's inherent ability to diffuse only small, lipophilic molecules. While strategies like carrier-mediated and receptor-mediated transcytosis offer targeted delivery capabilities, that are often saturable with limiting drug efficacy, and their development involves complex and costly modifications. Furthermore, increasing BBB permeability by means of chemical changes or sophisticated delivery methods, including targeting ligands or nanoparticles unintentionally raise production costs and provide difficulties for large-scale manufacture. During such procedures, there is also a chance that the protective integrity of the blood-brain barrier is compromised, exposing the brain to potentially harmful toxins and infections. Moreover, the production of generally efficient BBB-penetrating compounds is complicated by the BBB's individual diversity and its altered state in specific disorders. These drawbacks highlight the urgent need for innovative yet economical methods to raise the accuracy, security, and scalability of medicinal delivery to the brain. US8241670 describes the "Compositions capable of facilitating penetration across a biological barrier." The invention discloses a novel penetrating compositions including one or more effectors included within a water-soluble composition, immersed in a hydrophobic medium. The compositions are formulated for oral administrations that include biologically active effectors such as fibroblast growth factor (bFGF); hirudin, and hirulog. The formulation comprises fatty acid salts (e.g., sodium octanoate, sodium decanoate, sodium dodecanoate), a hydrophobic medium (e.g., aliphatic molecules, cyclic molecules, and aromatic molecules), a lecithin, a bile salt, or a non-ionic detergent, along with monoglycerides. Furthermore, the composition may incorporate pharmaceutically acceptable excipients or carriers. These compositions are suitable for encapsulation or tablet forms and enable effective delivery of therapeutic agents by improving bioavailability. The main drawback of this invention is manufacturing complexity due to the integration of multiple components, which raises formulation stability issues, production hurdles, high production costs, and low formulation efficiency. The production is manual, making it a time-consuming process. The invention does not mention artificial intelligence based tools for optimization of the formulation to reduce experimental efficiency.CN114642736A describes the "Blood-brain barrier penetrating drug delivery system and preparation method and application thereof." The invention discloses a blood-brain barrier penetrating drug delivery system comprises a neutrophil exosome and a drug encapsulated in the neutrophil exosome; the medicine is a medicine for treating central nervous system diseases. The invention constructs a neutrophil exosome drug-carrying delivery system with good biocompatibility. The drug-carrying system has biological characteristics similar to that of neutrophil, can efficiently penetrate through a blood-brain barrier, effectively respond to inflammation, targets an inflammation part, and is a drug carrier with application prospects for treating glioma and other central nervous system diseases. The invention provides a new strategy with a promising transformation application prospect for the non-invasive inflammatory microenvironment-targeted drug delivery of glioma and other central nervous system diseases. The main drawback of this invention is the complex preparation process due to the manual isolation and preparation of neutrophil exosomes, which is time-consuming, raises formulation stability issues, production hurdles, high production costs, and low formulation efficiency. The invention does not mention about artificial intelligence based tools for optimization of the formulation to reduce experimental efficiency.Therefore, there is a need of a hybridized gravitational search based system for planning blood-brain barrier penetrating molecule.OBJECT OF THE INVENTION The main object of the present invention is to provide a hybridized gravitational search based system for planning blood-brain barrier penetrating molecule.Another object of the present invention is to provide a system that includes explainable artificial intelligence modules for discovering and optimizing the peptide.Yet another object of the present invention is to provide a system that determines the characteristics to be optimized by hybridized gravitational search protocol along with critical amino acids in a peptide required for blood brain barrier penetration.Yet another object of the present invention is to provide a system that saves computational time and cost. Still another object of the present invention is to provide a system that utilizes population based search protocol, preferably hybridized gravitational search protocol to solve the issue of multiobjective optimization (MOO). SUMMARY OF THE INVENTIONThe present invention relates to a hybridized gravitational search based system for planning blood-brain barrier penetrating molecule that determines the characteristics to be optimized by hybridized gravitational search protocol along with critical amino acids in a peptide required for blood brain barrier penetration.In an embodiment, the present invention provides a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules. The system comprises a processor and a memory. The processor is configured to perform a hybridized gravitational search for planning blood brain barrier via creating, via a data module, a dataset by fetching data from one or more sources, computing, via feature extraction module, a set of handcrafted features corresponding to the dataset obtained from the data module and preparing an input data based on the set of handcrafted features, training, a first data model, via the set of handcrafted features for identification, via a machine learning classifier module of a set of prominent features according to a set of corresponding values, training, a second data model, via a set of sequence embedding for feature representation for identification, via a deep learning classifier module of one or more crucial amino acids and optimizing, via a gravitational search module to refine a search space and identify blood-brain barrier penetrating peptides (B3P2s).The above objects and advantages of the present invention will become apparent from the hereinafter set forth brief description of the drawings, detailed description of the invention, and claims appended herewith. BRIEF DESCRIPTION OF THE DRAWINGSAn understanding of the hybridized gravitational search based system for planning blood-brain barrier penetrating molecule of the present invention may be obtained by reference to the following drawings:Figure 1 is a block diagram of the hybridized gravitational search based system for planning blood-brain barrier penetrating molecule according to an embodiment of the present invention.Figure 2 is a schematic diagram of a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules according to an embodiment of the present invention.Figure 3 is a schematic architecture of explainable machine learning classifier (ML-B3P2pred) and explainable deep learning classifier (DL-B3P2pred) used in the hybridized gravitational search based system for planning blood-brain barrier penetrating molecules according to an embodiment of the present invention.Figure 4 is a schematic diagram of the particle encoding used in the hybridized gravitational search based system for planning blood-brain barrier penetrating molecules according to an embodiment of the present invention.Figure 5 is a flowchart depicting the working of present invention on a dataset comprising cell-penetrating peptides (CPPs), according to an embodiment of the present invention.Figure 6 is a graphical representation of an isometric mapping of the B3P2 peptide developed by hybridized gravitational search protocol, along with other cell-penetrating peptides (CPPs) and the peptides present in the dataset.Figure 7(a) to Figure 7(d) are graphical representations of a scatter plot of tradeoff achieved between objective values and the fitness values of all the particles.Figure 7(e) to Figure 7(j) are graphical representations of a pairwise relation between the values of objectives obtained by different particles.DETAILED DESCRIPTION OF THE INVENTION The present invention will now be described hereinafter with reference to the accompanying drawings in which a preferred embodiment of the invention is shown. This invention may, however, be embodied in many different forms and should not be construed as being limited to the embodiment set forth herein. Rather, the embodiment is provided so that this disclosure will be thorough, and will fully convey the scope of the invention to those skilled in the art.Many aspects of the invention can be better understood with references made to the drawings below. The components in the drawings are not necessarily drawn to scale. Instead, emphasis is placed upon clearly illustrating the components of the present invention. Moreover, like reference numerals designate corresponding parts through the several views in the drawings. Before explaining at least one embodiment of the invention, it is to be understood that the embodiments of the invention are not limited in their application to the details of construction and to the arrangement of the components set forth in the following description or illustrated in the drawings. The embodiments of the invention are capable of being practiced and carried out in various ways. In addition, the phraseology and terminology employed herein are for the purpose of description and should not be regarded as limiting. The present invention provides a hybridized gravitational search based system for planning blood-brain barrier penetrating molecule that determines the characteristics to be optimized by hybridized gravitational search protocol along with critical amino acids in a peptide required for blood brain barrier penetration.Referring to Figure 1, a block diagram of hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules is depicted. The hybridized gravitational search based system (100) for planning blood-brain comprises a processor (101) and a memory (102). The processor (101) is configured to perform a hybridized gravitational search for planning blood brain barrier via creating, via a data module (103), a dataset by fetching data from one or more sources. The processor (101) is configured to compute, via feature extraction module (104), a set of handcrafted features corresponding to the dataset obtained from the data module and preparing an input data based on the set of handcrafted features. The processor (101) is configured to train, a first data model, via the set of handcrafted features for identification, via a machine learning classifier module (105) of a set of prominent features according to a set of corresponding values The processor (101) is configured to train, a second data model, via a set of sequence embedding for feature representation for identification, via a deep learning classifier module (106) of one or more crucial amino acids. The processor (101) is configured to optimize, via a gravitational search module (106), to refine a search space and identify blood-brain barrier penetrating peptides (B3P2s).Further, the Figure 2 shows a schematic diagram of a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules according to an embodiment of the present invention.The objectives of the formulated fitness function have been determined by analyzing output of module (103). Also, crucial amino acids in a peptide that are not to be altered by HyGSA is determined by the analyzing the output.Further, Figure 3 elaborates the working of deep learning classifier (105), and machine learning classifier module (106). The machine learning classifier module (106) is trained via the set of hand crafted features obtained from the feature extraction module (104) and identify a set of prominent features from the set of hand crafted features according to a set of values of the set of hand crafted features. The first data model used herein is machine learning model which is preferably a ML-B3P2pred. The ML-B3P2 pred model used is made interpretable using explain-ability frameworks. For construction of the MLB3P2pred model, the XGB protocol is used, which is a supervised machine learning protocol based on the boosting technique. In the case of XGB, a set of decision trees is constructed iteratively, where each tree tries to correct the errors made by the previous tree. The 90% of the dataset is split into training and 10% of dataset to train and test the ML-B3P2pred model. Then, the XGB protocol is trained via fivefold cross-validation, and the optimal hyper-parameters are found using the Grid-Search CV protocol. After that, tools such as SHAP, Yellowbrick (using the Shapiro-Wilk protocol), and ELI5 are used to implement the model. The explainability tools are employed to find a set of features that are crucial in predicting a peptide as B3P2. Further, the deep learning classifier module (106) is trained via a set of sequence embedding for a feature representation and interpret a set of sequence based feature for protein classification. The second data model used herein is deep-learning based B3P2 model, which is preferably a DL-B3P2 pred. Moreover, the insights of both the first data model and first data model are combined for identifying one or more critical amino acid residues followed by applying an optimization with hybridized gravitational search protocol to refine a search space and identify the one or more critical amino acid residues with high potential for B3P2. The hybridized gravitational protocol refine the search space by initializing the population of gravitational search protocol using a set of peptides that are predicted as B3P2 by DL-B3P2 pred module, formulating a fitness / objective function using features found by the ML-B3P2 pred module and then utilizing the function to change the position of every particle, utilizing subset sum solution-based approach and measure of central tendency to determine the sample population for each iteration of gravitational search protocol, employing the DL-B3P2pred model to prevent important amino acid residues in a peptide / particle from changing in every iteration. Figure 4 depicts a schematic diagram of the particle encoding used in the hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules according to an embodiment of the present invention.Further, the steps include: Particles representation: The population consists of a set of peptides represented as particles. A dimension d in a particle Xi, denoted by xdi, corresponds to the amino acid at a given position d in a peptide pepi (represented as pepdi). Before encoding a given set of peptides as particles, the skip-gram protocol is utilized to find embedding vectors corresponding to each amino acid residue (20 natural amino acids). Then, to calculate xdi, the embedding vector for the amino acid (pepdi) has been averaged. Problem formulation: The hybridized gravitational search protocol then optimizes the four objectives which includes predicted probability value given by DLB3P2pred (PPV), the amino acid composition of tyrosine (AAC_Y), the normalized isoelectric point (npI), and the normalized molecular weight (nmw) for determining the blood brain barrier penetrating potential of given peptide.Further, after optimization, particles of the peptide are interacted via the hybridized gravitational search protocol to select a set of active particles from the set of peptides and fitness value is calculated of the particles of the peptide. The active particle selected are of low, average, and high fitness values that saves the time and computational cost. Following this, the cumulative fitness of the population is computed by summing up the fitness values of all the particles. After that, a random number is selected by selecting the one particle at a time from the number of particles of the peptides by generating a random number. Random selection of the particles ensures that the sample of the particles of peptides contains two-third of the number of particles that is required for penetration into the blood-brain barrier.Further, the particles are divided into random groups based on their fitness value that ensures a better tradeoff between the exploitation and exploration of the search space, thereby sorting the particles based on the fitness value into various candidate solutions. Then the candidate solutions are grouped into different non-dominated fronts, where each front contains solutions that have a fitness equivalent to all other solutions in the same front. Further, a final list of the B3P2 peptides is provided.EXAMPLE 1Experimental analysisReferring to Figure 5, a flow chart of the working of hybridized gravitational search based system (100) on a dataset comprising cell-penetrating peptides (CPPs) is depicted, according to an embodiment of the present invention. A set of experimentally validated and annotated cell-penetrating peptides (CPPs) of length 15 were collected. After that, the DL-B3P2pred model was used to find the peptides with the best chance to penetrate the BBB (with PPV > 0.9), and they were used to initialize the population of the hybridized gravitational search protocol. Then, hybridized gravitational search protocol was executed, wherein each particle's position and velocity were updated iteratively. After the protocol iterated a maximum number of times (k0), peptides from the first three fronts were given as output (considering Num = 5). Then, the isometric mapping (Isomap) technique was employed to map the distribution of the discovered B3P2s against the distribution of the peptides (B3P2 / non-B3P2) in the collected dataset.Referring to Figure 6, a graphical representation of an isometric mapping of the B3P2 peptide developed by hybridized gravitational search protocol, along with other cell-penetrating peptides (CPPs) and the peptides present in the dataset is depicted, according to an embodiment of the present invention. It was observed that the B3P2 peptides were found well-dispersed within the core region of experimentally validated B3P2 represents a particle of the population.Referring to Figure 7(a) to Figure 7(d), a graphical representation of a scatter plot of tradeoff achieved between objective values and the fitness values of all the particles are depicted according to an embodiment of the present invention. Referring to Figure 7(e) to Figure 7(j), a graphical representation of a pairwise relation between the values of objectives obtained by different particles are depicted, according to an embodiment of the present invention. Therefore, based on the Figure 7(a) to Figure 7(j), it is concluded that all the peptides in the first three fronts have the maximum possible value for predictability probability value (PPV) in the population. All the peptides in the first three fronts (except one) have amino acid composition of tyrosine (AAC_Y) values higher than the population's mean. All the peptides in the first three fronts have normalized isoelectric point (npI) values higher than the population's mean, with one peptide having a value almost equal to the maximum. All the peptides in the first three fronts have normalized molecular weight (nmw) values lower than the population's mean. Based on these observations, it is safe to say that the peptides developed for blood-brain barrier have required penetrable characteristics than other particles of the final population of the particles.Therefore, the present invention provides a hybridized gravitational search based system for planning blood-brain barrier penetrating molecules for discovering and optimizing the BBB penetrating peptides so that they have desirable characteristics such as high pI, low molecular weight, etc. Many modifications and other embodiments of the invention set forth herein will readily occur to one skilled in the art to which the invention pertain having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is to be understood that the invention is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation. The foregoing description of embodiments of the invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the invention. The embodiments were chosen and described in order to explain the principals of the invention and its practical application to enable one skilled in the art to utilize the invention in various embodiments and with various modifications as are suited to the particular use contemplated.
Claims
1. A hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules comprising: a processor (101); and a memory (102); wherein: said processor (101)is configured to perform a hybridized gravitational search for planning blood brain barrier via: i) creating, via a data module (103), a dataset by fetching data from one or more sources; ii) computing, via feature extraction module (104), a set of handcrafted features corresponding to the dataset obtained from the data module and preparing an input data based on the set of handcrafted features; iii) training, a first data model, via the set of handcrafted features for identification, via a machine learning classifier module (105) of a set of prominent features according to a set of corresponding values; iv) training, a second data model, via a set of sequence embedding for feature representation for identification, via a deep learning classifier module (106) of one or more crucial amino acids; v) optimizing, via a gravitational search module (106), to refine a search space and identify blood-brain barrier penetrating peptides (B3P2s).
2. The hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules as claimed in claim 1, wherein said first data model is machine learning module.
3. The hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules as claimed in claim 1, wherein said second data model is deep learning based B3P2 module.
4. The hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules as claimed in claim 1, wherein said machine learning model is ML-B3P2pred.
5. The hybridized gravitational search based system (100) for planning blood-brain barrier penetrating molecules as claimed in claim 1, wherein said deep learning based model is DL-B3P2pred. and described to explain the principles of the invention and its practical application to enable one skilled in the art to utilize the invention in various embodiments and with various modifications suited to the particular use contemplated.
