Survival prediction using methylomic profiles
AssigneeBioAge Labs, Ipc.
InventorsKristen Patricia Fortney, Jonah Daniel Sinick, Andrew Jarai Ho, Yonatan Nissan Donner
Life of this patent
Estimating the term…
Abstract
In various embodiments, the present description relates to the use of factors related to survival. The methods, compositions and systems described herein may be used to determine factors affecting survival, assess survival risk based on factors related to survival and/or make suggestions to increase the likelihood of survival or extend the period of survival.
Claims19
A computer system comprising a computer processor and a non-transitory computer readable storage medium storing instructions that, when executed by the processor, cause the processor to:
- access a training set of methylation fraction values for m methylation sites relating to j training subjects;
- train a machine-learned model using the accessed training set of methylation fraction values, the trained machine-learned model comprising a set of weights determined based on a classification of each methylation fraction value within the accessed training set of methylation fraction values indicating whether a sample associated with the methylation fraction value survived, the trained machine-learned model configured to predict one or more survival metrics based on methylation fraction values;
- access methylation fraction values for n nucleic acid methylation sites relating to a test subject;
- and
- apply the trained machine-learned model to the accessed methylation fraction values for the n nucleic acid methylation sites to generate a survival metric for the test subject based on the accessed methylation fraction values for the n nucleic acid methylation sites, wherein the survival metric represents a likelihood that the test subject will survive over an interval of time.
A method for predicting survival likelihood comprising:
- accessing a training set of methylation fraction values for m methylation sites relating to j training subjects;
- training a machine-learned model using the accessed training set of methylation fraction values, the trained machine-learned model comprising a set of weights determined based on a classification of each methylation fraction value within the accessed training set of methylation fraction values indicating whether a sample associated with the methylation fraction value survived, the trained machine-learned model configured to predict one or more survival metrics based on methylation fraction values;
- accessing methylation fraction values for n nucleic acid methylation sites relating to a test subject;
- and
- applying the trained machine-learned model to the accessed methylation fraction values for the n nucleic acid methylation sites to generate a survival metric for the test subject based on the accessed methylation fraction values for the n nucleic acid methylation sites, wherein the survival metric represents a likelihood that the test subject will survive over an interval of time.
Dependencies are inferred from the claim text.