202 to effectuate identification of candidate functional ingredients for blend formation and production and optimization thereof. While not illustrated separately, the analytic device 102 may comprise (or be communicatively connected to) a data aggregation service, e.g., that executes on the network 204.

Each of the compute device 202 and the analytic device 102 may be embodied as any type of compute device capable of performing functions, including, but not limited to, a compute device, a storage device, a server (e.g., stand-alone, rack-mounted, blade, etc.), a sled (e.g., a compute sled, an accelerator sled, a storage sled, etc.), an enhanced network interface controller (NIC), a network appliance (e.g., physical or virtual), a router, a web appliance, a distributed computing system, a processor-based system, and/or a multiprocessor system.

Referring now to FIGS. 3 and 4, at least one rank of a given functional ingredient 112 indicates a docking score of one or more active compounds 114 of that functional ingredient 112. A docking score may be indicative of a level of interaction, e.g., potency, between the one or more active compounds 114 of a particular functional ingredient 112 and a target receptor within an organism such as a human body. As just one example, a mechanism of action (MoA) of the one or more active compounds of a functional ingredient with a target receptor may indicate a binding mode and binding affinity of a complex formed by two or more constituent molecules with known structures.

With reference to FIG. 3, functional ingredients impacting sleep may act through a GABA pathway, e.g., by acting on GABAA receptors. The GABAA receptor 300 is a molecular target for suppressing activity of a central nervous system. The GABAA receptors include five subunits—two alpha (α) subunits 302, two beta (β) subunits 304, and one gamma (γ) subunit 308 arranged around a central pore (not illustrated). The GABA neurotransmitter (ligand) binds at two sites (GABA sites) 306 disposed between a and B subunits 302, 304 causing the central pore, also referred to as a chloride channel (Cl⁻ channel) of the GABAA receptor, to open. Benzodiazepines-like neurotransmitters bind at a benzodiazepine (BDZ) site (not illustrated) located between the α and γ subunit. Other binding sites, such as for zinc, furosemide, volatile anesthetics and/or alcohols, and convulsants, are also present.

For a GABAA receptor 300, most of active compounds dock at binding sites 302, 304 different from that of the GABA molecule that docks at site 306. The active compounds may cooperate with the GABA molecule to increase a period of time during which the GABAA receptor channel is open. The GABA molecule changes the intracellular chlorine (Cl⁻) concentration to enhance the protein function for sleep.

As one example, a docking score may range between 0 and 10. In other examples, different ranges and/or values are also contemplated. In some instances, the docking scoring mechanism and ranking may be set up such that a docking score having a larger magnitude indicates a stronger interaction or a higher activity interaction, than a docking score having a smaller magnitude. The bioactivities of at least one functional ingredient 112 in GABA pathway-based group of functional ingredients 112 may be ranked based on bioactivities (as expressed, for example, by docking scores) of active compounds 114 of the at least one functional ingredient 112 in the GABA pathway.

One of skill will appreciate that within an organism such as the human body, there are many biological pathways and that the reference in FIG. 3 to the GABA pathway merely provides an example of how the docking score may be obtained for any particular pathway.

With reference to FIG. 4, functional ingredients affecting sleep may act through Ado pathway, e.g., by affecting Ado levels, by acting on Ado receptors. Adenosine receptor 400 is one of the G protein-coupled receptors (GPCRs), also known as seven-transmembrane domain receptors or G protein-linked receptors (GPLR). The GPCR receptors are receptors that extend to a surface 406 of a cell to detect molecules outside (extracellular to) 402 the cell and activate cellular (intracellular) 404 responses. Ligands can bind either to extracellular N-terminus and loops (e.g., glutamate receptors) or to the binding site within transmembrane helices.

In the Ado pathway, active compounds with this MoA may be docked to adenosine A_2Aand A₁receptors, e.g., allosteric binding sites. Docking scores may be determined based on strength of binding between at least one active compound and the adenosine receptors. At least one functional ingredient 112 of adenosine pathway-based group of functional ingredients 112 may then be ranked based on the docking scores (such as, adenosine pathway docking scores) of the active compounds 114 present in the at least one functional ingredient 112.

FIG. 5 illustrates an exemplary implementation 500 for determining a reverse fingerprinting score by the reverse fingerprinting score module 108. Small molecules of active compounds can be represented using molecular descriptors, or molecular fingerprints, indicating physicochemical or structural features of molecules. Molecular fingerprints may be represented by bit strings, where each bit detects presence or absence of a specific chemical feature or represents a value range of a property descriptor.

Determining a reverse fingerprinting score includes evaluating bit patterns in active compounds being scored. The reverse fingerprinting score module 108 is configured to determine a reverse fingerprinting score by comparing a frequency of each bit position within the scored active compound and a frequency of the position within bit strings of reference molecules. For example, assigned bits may be indicative of features of the molecule such as functional groups, pharmacophore triangles and concentric circles. A reverse fingerprint score indicates a probability (likelihood) of a given active compound to perform specific function.

In one example, the reverse fingerprinting score module 108 includes a training data database 502, small molecule structures 504, and a reverse fingerprinting model 506. The training data database 502 includes digital molecular fingerprint data generated based on the molecular structure of several compounds. The reverse fingerprinting model 506 may be trained using the training data set 502. In addition to the training data, the reverse fingerprinting model 506 is configured to receive, as input, the small molecule structures 504 of the one or more active compounds 114. The reverse fingerprinting model 506 is configured to output a probability value, which may be used to rank the ingredients.

The reverse fingerprinting module 108 scores the activity of at least one active compound for virtual screening, narrowing down and ranking the functional ingredients to be chosen for a blend combination, and predicting the activity of new functional ingredients. The docking score and the reverse fingerprinting score indicate an expected activity of a particular active compound. Moreover, both scores may be based on the performance or efficacy of a particular active compound within an identified pathway. Under the method of the present disclosure the same active compound may have a different score for at least one of differing pathways.

With reference to FIGS. 6A-6F, the multi-parameter optimization score module 110 determines a membrane barrier score of one or more of the active compounds 114. A membrane barrier score of the one or more active compounds 114 of a given functional ingredient 112 indicates the capability of that active compound to cross the blood-brain membrane barrier. A multi-parameter optimization score is indicative of the ability of the active compound to cross the blood-brain membrane barrier to reach the central nervous system. FIGS. 6A-6F illustrate exemplary piecewise linear transformational functions with values between 0 and 1. Each transformational function indicates a range of values spanned by a given physicochemical property of a compound, including the most and least desirable values from an optimization standpoint.

FIG. 6A illustrates a range of values of lipophilicity (C log P). FIG. 6B illustrates a calculated distribution coefficient at pH 7.4 (C log D). FIG. 6C illustrates the molecular weight (MW) of the active compound. FIG. 6D illustrates the topological polar surface area (PSA) of the active compound. FIG. 6E illustrates a range of a number of hydrogen-bond donors (HBD). FIG. 6F illustrates the acid-base dissociation constant (pKa) value range of a most basic center molecule of the compound.

The multi-parameter optimization score module 110 may eliminate a functional ingredient 112 from among the potential candidate functional ingredients if one or more of its active compounds present in the corresponding potential candidate functional ingredient has a blood-brain membrane barrier score greater than or less than a predefined value. A blood-brain membrane barrier score may be determined by molecular structure properties of the one or more active compounds. The blood-brain membrane barrier score may be independent of the activity of the one or more active compounds within any particular biological pathway.

For example, ranking and selecting based on at least one mechanism of action and pathway may be indicative of top (i.e., more preferred) functional ingredient candidates. The identified functional ingredient candidates may then be evaluated based on one or more other factors, such as, but not limited to, dosage, safety, and stability. The systems and methods of the present disclosure enable identifying high efficacy functional ingredient blends using categorization and ranking of active compounds of individual functional ingredients.

FIG. 7A illustrates an exemplary process 700-A for generating functional ingredient blends in accordance with the present disclosure. In some embodiments, the process 700-A may be executed by one or more processors of the analytic device 102 using data sources and/or one or more modules of the analytic device 102 (e.g., the functional ingredient and biological pathway database 120, the pathway-based group receiving module 104, the docking score module 106, the reverse fingerprinting score module 108, and the multi-parameter optimization score module 110). The process 700-A may begin at block 702 where the analytic device 102 receives, e.g., from the input data from the data sources 206 such as the functional ingredients and biological pathway database 120, a plurality groups of functional ingredients 112 as potential candidates for a combination in a functional ingredient blend. In some instances, at least one group combines the plurality of functional ingredients 112 that use the same biological pathway to achieve one or more desired effects of the functional ingredient 112 on an organism, such as to activate, inhibit, increase, decrease, modulate, or otherwise affect a given biological process or function of an organism.

At block 704, the analytic device 102 determines at least one of a plurality of scores for one or more of the active compounds 114 of at least one of the plurality of functional ingredients 112 of at least one pathway-based group of functional ingredients 112. As described in reference to at least FIG. 7B, the analytic device 102 may determine, for the at least one active compound 114, at least one of a docking score, a reverse fingerprinting score, and a multi-parameter optimization score. At block 706, the analytic device 102 ranks the at least one functional ingredient 112 based on the at least one of the plurality of scores of the at least one active compound 114 of the at least one functional ingredient 112. The analytic device 102, at block 708, selects, from at least one pathway-based group, one or more functional ingredients 112 having a rank greater than a threshold, where the selected functional ingredients are candidate ingredients for combination in functional ingredient blends.

The process 700-A may then end. In other examples, the process 700-A may be repeated in response to a receipt, by the analytic device 102, of a plurality of pathway-based groups, each group including one functional ingredient 112 or more than one functional ingredient 112, i.e., a plurality of functional ingredients 112, or in response to a receipt of different data or signal.

FIG. 7B illustrates an exemplary process 700-B for determining scores of active compounds 114 for identifying candidate functional ingredient for combination in functional ingredient blends. One or more operations of the process 700-B may be performed by the analytic device 102 at block 704 described in reference to FIG. 7A. The process 700-B begins at block 716 where the analytic device 102 selects a first functional ingredient of a plurality of functional ingredients 112 within a given pathway-based group in accordance with the present disclosure.

The analytic device 102, at block 718, determines at least one of a docking score, a reverse fingerprinting score, and a multi-parameter optimization score of at least one of the plurality of active compounds 114. In one example, a docking score of a given active compound 114 indicates a level of interaction, e.g., potency, between that active compound 114 and a target receptor of an organism such as a human body.

In another example, to determine a reverse fingerprinting score of the one or more active compounds 114 of the selected functional ingredient 112, the analytic device 102 compares a bit string of the active compound being scored (scored active compound) to bit strings of reference molecules. In particular, the analytic device 102 may compare a frequency of each bit position within the scored active compound to the frequency of the position in bit strings of the reference molecules.

In still another example, the analytic device 102 determines the multi-parameter optimization score of a given active compound 114 by evaluating how quickly and/or how completely that active compound crosses a blood-brain membrane barrier. At block 720, the analytic device 102 determines ranks of the selected functional ingredient 112 based on at least one score of the one or more of the active compounds 114 of that functional ingredient 112.

The analytic device 102, at block 722, determines whether the rank of the selected functional ingredient is greater than a predefined threshold. In response to the rank of the selected functional ingredient 112 being less than a threshold, the analytic device 102, at block 724, discards the selected functional ingredient 112 as a candidate ingredient 122 for combination in a functional ingredient blend.

In response to the selected functional ingredient 112 being greater than a threshold, the analytic device 102, at block 726, adds the selected functional ingredient 112 as a candidate ingredient 122 for generating a functional ingredient blend. The process 700-B may then proceed to block 728 where the analytic device 102 determines whether more functional ingredients 112 are desired to be scored. If another functional ingredient is desired to be scored, the process 700-B returns to block 716 where the analytic device 102 selects a next functional ingredient 112. If no other functional ingredients remain to be scored within a given pathway-based group, the analytic device 102 may exit the process 700-B.

The process 700-B may then end. In other examples, the process 700-B may be repeated in response to a determination, by the analytic device 102, that a plurality of functional ingredients 112 in several pathway-based groups may be tested as candidates for combination in a functional ingredient blend. The process 700-B may also be repeated in response to a different determination or signal.

FIG. 8 illustrates an exemplary implementation 800 of the analytic device 102. While the illustrated implementation 800 describes only the analytic device 102, in other examples, the compute device 202 may be embodied to include similar components configured to perform similar operations to those described with respect to the analytic device 102. The analytic device 102 includes an analytic compute engine 802, an I/O subsystem 808, one or more data storage devices 810, and communication circuitry 812. It will be appreciated that the analytic device 102 may include other or additional components, such as those commonly found in a typical computing device (e.g., various input/output devices and/or other components), in other embodiments. Additionally, in some embodiments, one or more of the illustrative components may be incorporated in, or otherwise form a portion of, another component.

The analytic compute engine 802 may be embodied as any type of device or collection of devices capable of performing the described various compute functions, such as, but not limited to, compute functions related to or bearing on one or more of receiving a plurality of groups of functional ingredients, functional ingredients of a given group of functional ingredients affecting a same predefined organism function and each functional ingredient including a plurality of active compounds associated with that functional ingredient, determining one or more scores of at least one active compound of at least one functional ingredient of at least one pathway-based group of functional ingredients, determining one or more ranks of at least one functional ingredient based on the one or more scores of at least one active compound of the at least one functional ingredient, and generating a blend of two or more functional ingredients from at least one functional ingredient of at least one pathway-based group of functional ingredients having one or more ranks greater than a threshold.

In some embodiments, the analytic compute engine 802 may be embodied as a single device, such as an integrated circuit, an embedded system, a field-programmable gate array (FPGA), a system-on-a-chip (SOC), an application-specific integrated circuit (ASIC), reconfigurable hardware or hardware circuitry, or other specialized hardware to facilitate performance of the functions described herein. In some embodiments, the analytic compute engine 802 may include, or may be embodied as, one or more processors 804 (i.e., one or more central processing units (CPUs)) and memory 806.

The processor(s) 804 may be embodied as any type of processor capable of performing the described functions, such as, but not limited to, functions related to or bearing on receiving, processing, or otherwise manipulating, and/or transmitting data to facilitate one or more of receiving at least one group of functional ingredients, where a group may include one functional ingredient or more than one functional ingredient, where each functional ingredient may include one or more active compounds associated with that functional ingredient, and where each functional ingredient of a given group of functional ingredients affects a predefined organism function using one and/or same biological pathway, determining one or more scores of at least one active compound of at least one functional ingredient of at least one pathway-based group of functional ingredients, determining one or more ranks of at least one functional ingredient based on the one or more scores of at least one active compound of that functional ingredient, and generating a blend of two or more functional ingredients from functional ingredients having one or more ranks greater than a threshold. For example, the processor(s) 804 may be embodied as one or more single-core processors, one or more multi-core processors, a digital signal processor, a microcontroller, or other processor or processing/controlling circuit(s). In some embodiments, the processor(s) 804 may be embodied as, include, or otherwise be coupled to an FPGA, an ASIC, reconfigurable hardware or hardware circuitry, or other specialized hardware to facilitate performance of the described functions.

The memory 806 may be embodied as any type of volatile (e.g., dynamic random access memory (DRAM), etc.) or non-volatile memory or data storage capable of performing the described functions, such as, but not limited to, functions related to or bearing on storing, writing, erasing, overwriting, aggregating, buffering, or otherwise manipulating data to facilitate one or more of receiving a plurality of groups of functional ingredients, where each group includes one or more functional ingredients and each functional ingredient includes one or more active compounds associated with that functional ingredient, and where one or more functional ingredients of a given group affects a predefined organism function using a same corresponding biological pathway, determining one or more scores of at least one active compound of at least one functional ingredient of at least one pathway-based group of functional ingredients, determining one or more ranks of at least one functional ingredient based on the one or more scores of at least one active compound of that functional ingredient, and generating a blend of two or more functional ingredients by combining at least one functional ingredient having one or more ranks greater than a threshold. It will be appreciated that the memory 806 may include main memory (i.e., a primary memory) and/or cache memory (i.e., memory that can be accessed more quickly than the main memory). Volatile memory may be a storage medium that requires power to maintain the state of data stored by the medium. Non-limiting examples of volatile memory may include various types of random access memory (RAM), such as DRAM or static random access memory (SRAM).