Marker screening and signal detection

A computer-implemented method for increasing the likelihood of an actionable result from whole genome sequencing or whole exome sequencing (WGS/WES) testing conducted on a human patient, using a computer program operating on a computing device, the method comprising:

• receiving electronic health record (EHR) data for the human patient, the EHR data including information indicating phenotypic features of the human patient;
• encoding, using natural language processing, a set of clinical codes from the phenotypic features included in the EHR data, the set of clinical codes encoded such that a particular clinical code is a machine readable representation of at least one phenotypic feature but any particular phenotypic feature is represented by no more than one machine readable clinical code;
• assembling a clinical code table from the set of clinical codes comprising each clinical code of the set of clinical codes and each combination of at least two clinical codes of the set of clinical codes;
• accessing a marker code combination set, stored in memory, comprising a plurality of marker clinical-code combinations each corresponding to a positive diagnoses from WGS/WES testing, the plurality of marker clinical-code combinations generated from output of a trained Bayesian confidence-propagation neural network, or Multi-item Gamma Poisson Shrinker;
• identifying at least one clinical code or combination of at least two clinical codes from the clinical code table that is present in the marker code combination set;
• responsive to the identifying clinical code or combination of at least two clinical codes from the clinical code table, generating a recommendation to perform WGS/WES testing on the human patient;
• and
• providing the recommendation as a visual alert within a user interface associated with a remote computing device that is communicatively coupled with the EHR.

One or more computer-readable media having computer-usable instructions embodied thereon that, when executed by a computer processor, enable the computer processor to perform a method for generating a recommendation to conduct whole genome sequencing or whole exome sequencing (WGS/WES) testing on a human patient based on a likelihood of receiving an actionable diagnosis, the method comprising:

• receiving electronic health record (EHR) data for a plurality of patients, the EHR data for each patient of the plurality of patients including phenotypic features, an indication that WGS/WES testing was completed, and a diagnostic indicator that indicates whether a diagnoses was made or confirmed based on the results of the WGS/WES testing;
• generating a set of clinical codes representative of the phenotypic features, the set of clinical codes generated such that a particular clinical code is representative of at least one phenotypic feature but any particular phenotypic feature is represented by no more than one clinical code;
• assembling a clinical code table for each patient of the plurality of patients comprising each clinical code that corresponds to the phenotypic features included in the EHR data for the patient;
• encoding machine readable code combination sets including each combination of at least a pair of clinical codes from the clinical code table for each patient of the plurality of patients;
• computing using a Bayesian confidence-propagation neural network or Multi-item Gamma Poisson Shrinker mark code combinations having a probability greater than a predetermined threshold from the machine readable code combination sets, the Bayesian confidence-propagation neural network or Multi-item Gamma Poisson Shrinker trained to determine the probability a code combination set corresponds to a positive diagnostic indicator;
• generating a recommendation for administration of a WGS/WES test in a EHR for a patient not included in the plurality of patients utilizing the marker code combinations as an automatic filter to identify the patient based on the EHR for the patient including at least one phenotypic feature combination corresponding to at least one marker code combination of marker code combinations;
• and
• providing the recommendation over a user interface associated with a remote computing device that is communicatively coupled with the EHR.

One or more computer-readable media having computer-usable instructions embodied thereon that, when executed by a computer processor, enable the computer processor to perform a method for increasing the likelihood of an actionable result from whole genome sequencing or whole exome sequencing (WGS/WES) testing conducted on a human patient, using a computer program operating on a computing device, the method comprising:

• receiving electronic health record (EHR) data for the human patient, the EHR data including information indicating phenotypic features of the human patient;
• encoding, using natural language processing, a set of clinical codes from the phenotypic features included in the EHR data, the set of clinical codes encoded such that a particular clinical code is a machine readable representation of at least one phenotypic feature but any particular phenotypic feature is represented by no more than one machine readable clinical code;
• assembling a clinical code table from the set of clinical codes comprising each clinical code of the set of clinical codes and each combination of at least two clinical codes of the set of clinical codes;
• accessing a marker code combination set, stored in memory, comprising a plurality of marker clinical-code combinations each corresponding to a positive diagnoses from WGS/WES testing, the plurality of marker clinical-code combinations generated from output of a trained Bayesian confidence-propagation neural network, or Multi-item Gamma Poisson Shrinker;
• identifying at least one clinical code or combination of at least two clinical codes from the clinical code table that is present in the marker code combination set;
• responsive to the identifying clinical code or combination of at least two clinical codes from the clinical code table, generating a recommendation to perform WGS/WES testing on the human patient;
• and
• providing the recommendation as a visual alert within a user interface associated with a remote computing device that is communicatively coupled with the EHR.