# US 2022/0020453 A1 · Preemptible-based scaffold hopping

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0020453 A1
- Document ID: US20220020453A1
- Status: Published application
- Kind: A1
- Application number: 16/933,225
- Filed: 2020-07-20
- Published: 2022-01-20
- Earliest claimed: 2020-07-20
- Assignees: Recursion Pharmaceuticals, Inc.
- Applicants: Recursion Pharmaceuticals, Inc.
- Inventors: Mason Victors; Nathan Wilkinson; Scott Nielsen; Jorge Aguilera-Iparraguirre
- Web page: https://patentgrep.com/patent/US20220020453A1

## CPC

- G06F 11/1479: Generic software techniques for error detection or fault masking
- G16C 20/40: Searching chemical structures or physicochemical data
- G06F 9/4881: {Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues}
- G06F 9/54: Interprogram communication
- G06F 11/0715: {in a system implementing multitasking (multitasking per se G06F9/46)}
- G16C 20/90: Programming languages; Computing architectures; Database systems; Data warehousing
- G16C 20/50: Molecular design, e.g. of drugs
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G16C: COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE

## Abstract

Method and apparatus for molecular scaffold hopping. Sending instructions, from an interface of a scheduler computer to a job runner computer, the instructions prepared by a processor of the scheduler computer, to perform a plurality of separate computational tasks, wherein each of the separate computational tasks includes calculating one or more chemical properties for a query molecule or for molecules in a library of molecules, wherein one or more of the plurality of separate computational tasks performed on the job runner computer are preemptible computing instances. Receiving status indicators, by the interface, the status indicators sent from the job runner computer, for each of the plurality of separate computational tasks, the indicators are one of: incomplete, completed or failed computing instances. Resending to the job runner computer by the interface, the instructions to the job runner computer that correspond to the separate computational tasks having the failed computing instance indicator to thereby increase fault-tolerance against the separate computational tasks not attaining the completed computing instance indicator.

## Claims

### Claim 1

A method of molecular scaffold hopping, comprising:
sending instructions, from an interface of a scheduler computer to a job runner computer, the instructions prepared by a processor of the scheduler computer, to perform a plurality of separate computational tasks, wherein each of the separate computational tasks includes calculating one or more chemical properties for a query molecule or for molecules in a library of molecules, wherein one or more of the plurality of separate computational tasks performed on the job runner computer are preemptible computing instances;
receiving status indicators, by the interface, the status indicators sent from the job runner computer, for each of the plurality of separate computational tasks, wherein each of the status indicators are one of: an incomplete computing instance, a completed computing instance or a failed computing instance;
resending to the job runner computer by the interface, the instructions to the job runner computer that correspond to the separate computational tasks having the one status indicator of the failed computing instance to thereby increase fault-tolerance against the one or more of the plurality of separate computational tasks not attaining the status indicator of the completed computing instance.

### Claim 2 (depends on 1)

The method of claim 1, wherein the resending of the instructions does not include the instructions corresponding to the separate computational tasks having the status indicator equal to the incomplete computing instance or equal to the complete computing instance.

### Claim 3 (depends on 1)

The method of claim 1, wherein the resending of the instructions further includes an iteration instruction to repeat the resending of the instruction for a plurality of times for the separate computational tasks continuing to have the one status indicator of the failed computing instance.

### Claim 4 (depends on 1)

The method of claim 1, wherein all of the separate computational tasks performed on the job runner computer are preemptible computing instances.

### Claim 5 (depends on 1)

The method of claim 1, wherein the failed computing instance occurs due to one or more of:
a preemption of at least one of the separate computational tasks;
a network partition between the scheduler computer and the job runner computer;
a network partition between the job runner computer and a resource computer that includes a digital record of chemical information about the query molecule or the molecules in the library of molecules; or
an excess number of information requests from the job runner computer to a resource computer that includes a digital record of chemical information about the query molecule or the molecules in the library of molecules.

### Claim 6 (depends on 1)

The method of claim 1, wherein the instructions sent from the interface of the scheduler computer to the job runner computer include instructions to divide the plurality of separate computational tasks into portions that are completed in an average completion time that is less than an average preemption time for the one or more of the separate computational instances.

### Claim 7 (depends on 1)

The method of claim 1, wherein the instructions sent from the interface of the scheduler computer to the job runner computer include instructions to initiate performing at least two of the separate computational tasks at different times from each other

### Claim 8 (depends on 1)

The method of claim 1, wherein the separate computational tasks to calculate the one or more chemical properties includes molecular weight filtering of the molecules in the library of molecules to select a subset library of molecules that are within a multiplier of a molecular weight of the query molecule,

### Claim 9 (depends on 1)

The method of claim 1, wherein the separate computational tasks to calculate the one or more chemical properties includes calculating an ensemble of 3D conformers of the query molecule or of the molecules in the library.

### Claim 10 (depends on 1)

The method of claim 1, wherein the separate computational tasks to calculate the one or more chemical properties includes calculations to align one of the 3D conformers of the query molecule with one of the 3D conformers of one of the molecules in the library of molecules to form an aligned conformer pair having a minimized molecular volumes difference and calculating a Tanimoto shape similarity score representative of the minimized molecular volumes difference.

### Claim 11 (depends on 1)

The method of claim 1, wherein the separate computational tasks to calculate the one or more chemical properties includes calculating a 2D molecular fingerprint difference between the query molecule and one of the molecules of the library of molecules, and, calculating a Tanimoto molecular fingerprint similarity score representative of the 2D molecular fingerprint difference

### Claim 12

An apparatus for molecular scaffold hopping, the apparatus comprising:
a scheduler computer, the scheduler computer including a processor and an interface, wherein:
the processor is configured to prepare instructions to perform a plurality of separate computational tasks, wherein each of the separate computational tasks includes calculating one or more chemical properties for a query molecule or for molecules in a library of molecules, wherein one or more of the plurality of separate computational tasks are preemptible computing instances, and
the interface is configured to:
send the instructions to a job runner computer,
receive from the job runner computer, a status indicator of each of the plurality of separate computational tasks, wherein the status indicator is one of an incomplete computing instance, a completed computing instance or a failed computing instance, and
resend to the job runner computer, the instructions that correspond to the separate computational tasks having the status indicator of the failed computing instance to thereby increase fault-tolerance against the one or more of the plurality of separate computational tasks not attaining the status indicator of the completed computing instance.

### Claim 13 (depends on 12)

The apparatus of claim 12, wherein the apparatus is part of a computing system that also includes the job runner computer.

### Claim 14 (depends on 12)

The apparatus of claim 12, wherein the instructions prepared by the processor and sent by the interface includes an iteration instruction to repeat the resending of the instruction for a plurality of times for the separate computational tasks continuing to have the one status indicator of the failed computing instance.

### Claim 15 (depends on 12)

The apparatus of claim 12, wherein the processor is configured to prepare the instructions based on dividing the plurality of separate computational tasks into portions that are completed in an average completion time that is less than an average preemption time for the one or more of the separate computational instances.

### Claim 16 (depends on 12)

The apparatus of claim 12, wherein the instructions prepared by the processor and sent by the interface include instructions initiate performing at least two of the separate computational tasks at different times from each other.

### Claim 17 (depends on 12)

The apparatus of claim 12, wherein at least one the separate computational tasks to calculate the one or more chemical properties includes molecular weight filtering of the molecules in the library of molecules to select a subset library of molecules that are within a multiplier of a molecular weight of the query molecule.

### Claim 18 (depends on 12)

The apparatus of claim 12, wherein at least one of the separate computational tasks to calculate the one or more chemical properties includes calculating an ensemble of 3D conformers of the query molecule or of the molecules in the library.

### Claim 19 (depends on 12)

The apparatus of claim 12, wherein at least one the separate computational tasks to calculate the one or more chemical properties includes calculations to align one of the 3D conformers of the query molecule with one of the 3D conformers of one of the molecules in the library of molecules to form an aligned conformer pair having a minimized molecular volumes difference and calculating a Tanimoto shape similarity score representative of the minimized molecular volumes difference.

### Claim 20 (depends on 12)

The apparatus of claim 12, wherein at least one of the separate computational tasks to calculate the one or more chemical properties includes calculating a 2D molecular fingerprint difference between the query molecule and one of the molecules of the library of molecules, and, calculating a Tanimoto molecular fingerprint similarity score representative of the 2D molecular fingerprint difference.
