Method, apparatus and computer program product for providing a multi-omics framework for estimating temporal disease trajectories

Patent metadata

Field	Value
Publication number	US-2025299831-A1
Application number	US-202519229258-A
Country	US
Kind code	A1
Filing date	Jun 5, 2025
Priority date	Jun 14, 2021
Publication date	Sep 25, 2025
Grant date	—

Abstract

Methods, apparatuses, systems, computing devices, computing entities, and/or the like are provided. An example method may include selecting at least one client profile data object from a plurality of client profile data objects; retrieving at least one initial transcriptome data object and at least one subsequent transcriptome data object associated with the at least one client profile data object; generating at least one dynamic multigraph data object based at least in part on the at least one initial transcriptome data object, the at least one subsequent transcriptome data object, and at least one clinical event data object; training a temporal graph network based at least in part on the at least one dynamic multigraph data object to generate a risk window prediction data object; and performing at least one data operation based at least in part on the risk window prediction data object.

First claim

1 . A system comprising:
one or more processors; and one or more memories storing processor-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising: receiving, via a user interface, an input that identifies a client profile associated with a transcriptome data object and a disease identifier, wherein the transcriptome data object is associated with a clinical event; receiving a graph representation comprising a plurality of nodes and a plurality of edges associated with the disease identifier; storing the transcriptome data object as one or more nodes within the graph representation, wherein the graph representation comprises one or more time axes that are associated with a temporal progression of a disease corresponding to the disease identifier and each of the one or more nodes is linked to at least one of one or more temporal identifiers positioned on the one or more time axes; generating, using a temporal graph network (TGN) and based at least in part on the graph representation, a risk window prediction that comprises an estimated lower bound metadata and an estimated upper bound metadata associated with the disease identifier, wherein the TGN is trained by:
comparing a training risk window prediction with a training onset temporal metadata for the training risk window prediction, and
responsive to the training onset temporal metadata falling outside of the training risk window prediction, adjusting one or more parameters of the TGN and regenerating the training risk window prediction; and
rendering, using the user interface, the risk window prediction.

2 . The system of claim 1 , wherein the TGN is further trained by:
receiving validated onset temporal metadata for the risk window prediction; and responsive to the validated onset temporal metadata falling outside of the risk window prediction, adjusting the one or more parameters of the TGN.

3 . The system of claim 1 , wherein the operations further comprise:
generating an updated graph representation from the graph representation responsive to an additional clinical event or an additional transcriptome data object for the client profile; and regenerating the risk window prediction based at least in part on the updated graph representation.

4 . The system of claim 1 , wherein the transcriptome data object is associated with an initial temporal identifier, and the operations further comprise:
receiving a subsequent transcriptome data object associated with a subsequent temporal identifier; and generating the graph representation based at least in part on the initial temporal identifier, the subsequent transcriptome data object, and the subsequent temporal identifier.

5 . The system of claim 4 , wherein the operations further comprise:
calculating differential expression metadata based at least in part on the transcriptome data object and the subsequent transcriptome data object; and generating the graph representation based at least in part on the differential expression metadata.

6 . The system of claim 4 , wherein the initial temporal identifier corresponds to the clinical event and the subsequent temporal identifier corresponds to a subsequent clinical event that is subsequent to the clinical event.

7 . The system of claim 4 , wherein the client profile is associated with a whole-genome sequence (WGS) data object and the transcriptome data object and the WGS data object are associated with the initial temporal identifier.

8 . The system of claim 7 , wherein the WGS data object comprises at least one of polygenic risk score (PRS) metadata related to the disease identifier or combined PRS and phenome-wide association study (PRS-PheWAS) metadata related to the disease identifier.

9 . The system of claim 1 , wherein the transcriptome data object comprises transcriptome metadata related to a particular tissue or cell associated with the disease identifier.

10 . The system of claim 9 , wherein the transcriptome data object comprises an initial single-cell ribonucleic acid (RNA) sequencing assay (scRNA-seq) metadata associated with the disease identifier.

11 . A computer-implemented method comprising:
receiving, by one or more processors and via a user interface, an input that identifies a client profile associated with a transcriptome data object and a disease identifier, wherein the transcriptome data object is associated with a clinical event; receiving, by the one or more processors, a graph representation comprising a plurality of nodes and a plurality of edges associated with the disease identifier; storing, by the one or more processors, the transcriptome data object as one or more nodes within the graph representation, wherein the graph representation comprises one or more time axes that are associated with a temporal progression of a disease corresponding to the disease identifier and each of the one or more nodes is linked to at least one of one or more temporal identifiers positioned on the one or more time axes; generating, by the one or more processors, using a temporal graph network (TGN), and based at least in part on the graph representation, a risk window prediction that comprises an estimated lower bound metadata and an estimated upper bound metadata associated with the disease identifier, wherein the TGN is trained by:
comparing a training risk window prediction with a training onset temporal metadata for the training risk window prediction, and
responsive to the training onset temporal metadata falling outside of the training risk window prediction, adjusting one or more parameters of the TGN and regenerating the training risk window prediction; and
rendering, by the one or more processors and using the user interface, the risk window prediction.

12 . The computer-implemented method of claim 11 , wherein the TGN is further trained by:
receiving validated onset temporal metadata for the risk window prediction; and responsive to the validated onset temporal metadata falling outside of the risk window prediction, adjusting the one or more parameters of the TGN.

13 . The computer-implemented method of claim 11 , further comprising:
generating an updated graph representation from the graph representation responsive to an additional clinical event or an additional transcriptome data object for the client profile; and regenerating the risk window prediction based at least in part on the updated graph representation.

14 . The computer-implemented method of claim 11 , wherein the transcriptome data object is associated with an initial temporal identifier, and the computer-implemented method further comprises:
receiving a subsequent transcriptome data object associated with a subsequent temporal identifier; and generating the graph representation based at least in part on the initial temporal identifier, the subsequent transcriptome data object, and the subsequent temporal identifier.

15 . The computer-implemented method of claim 14 , further comprising:
calculating differential expression metadata based at least in part on the transcriptome data object and the subsequent transcriptome data object; and generating the graph representation based at least in part on the differential expression metadata.

16 . The computer-implemented method of claim 14 , wherein the initial temporal identifier corresponds to the clinical event and the subsequent temporal identifier corresponds to a subsequent clinical event that is subsequent to the clinical event.

17 . The computer-implemen

Assignees

• Optum Services Ireland Ltd

Inventors

• Godden Paul J

Classifications

• G16H10/60

  for patient-specific data, e.g. for electronic patient records · CPC title

• G16B25/10

  Gene or protein expression profiling; Expression-ratio estimation or normalisation · CPC title

• G16B50/30

  Data warehousing; Computing architectures · CPC title

• G16H10/40

  for data related to laboratory analysis, e.g. patient specimen analysis · CPC title

• G16B40/20

  Supervised data analysis · CPC title