Predicting likelihood and site of metastasis from patient records

A method for predicting metastasis of a cancer in a subject, comprising:

• at a computer system having one or more processors, and memory storing one or more programs for execution by the one or more processors:
• (A) obtaining, in electronic format, for each respective training subject in a plurality of training subjects, a respective plurality of data elements, wherein:
• each respective training subject in the plurality of training subjects has a corresponding cancer, and
• for each respective training subject in the plurality of training subjects, the corresponding plurality of data elements comprises (i) a first corresponding set of sequence features comprising relative abundance values for the expression of a plurality of at least 30 genes in a sample of the respective training subject's cancer, and (ii) a corresponding indication of whether the respective training subject's cancer metastasized;
• (B) training one or more predictive models using a set of independent variables and a set of dependent variables for the training, wherein:
• the set of independent variables comprises, for each respective training subject in the plurality of training subjects, the first respective set of sequences features comprising the relative abundance values for the expression of the plurality of at least 30 genes, and
• the set of dependent variables comprises, for each respective training subject in the plurality of training subjects, the corresponding indication of whether the respective subject's cancer metastasized,
• thereby obtaining one or more trained models;
• and
• (C) using the one or more trained models to collectively provide a respective one or more indications of whether a cancer will metastasize in a test subject, thereby predicting whether the cancer will metastasize.

A computer system having one or more processors, and memory storing one or more programs for execution by the one or more processors, the one or more programs comprising instructions for performing a method for predicting metastasis of a colorectal cancer in a subject, the method comprising:

• (A) obtaining a plurality of at least 10,000 sequence reads, wherein the plurality of sequence reads is obtained for a plurality of RNA molecules from a sample of the cancer obtained from the subject;
• (B) determining, from the plurality of at least 10,000 sequence reads, a plurality of data elements for the subject's cancer comprising:
• a first set of sequence features comprising relative abundance values for the expression of a plurality of at least 30 genes in the sample of the cancer obtained from the subject, and
• (C) applying, to the plurality of data elements for the subject's cancer comprising the relative abundance values for the expression of the plurality of at least 30 genes, one or more models that are collectively trained to provide a respective one or more indications of whether the cancer will metastasize in the subject, thereby predicting whether the cancer will metastasize.

A non-transitory computer readable storage medium storing one or more programs configured for execution by a computer, the one or more programs comprising instructions that, when executed by a processor, cause the processor to perform a method for predicting metastasis of a colorectal cancer in a subject, the method comprising:

• (A) obtaining a plurality of at least 10,000 sequence reads, wherein the plurality of sequence reads is obtained for a plurality of RNA molecules from a sample of the cancer obtained from the subject;
• (B) determining, from the plurality of at least 10,000 sequence reads, a plurality of data elements for the subject's cancer comprising:
• a first set of sequence features comprising relative abundance values for the expression of a plurality of at least 30 genes in the sample of the cancer obtained from the subject, and
• (C) applying, to the plurality of data elements for the subject's cancer comprising the relative abundance values for the expression of the plurality of at least 30 genes, one or more models that are collectively trained to provide a respective one or more indications of whether the cancer will metastasize in the subject, thereby predicting whether the cancer will metastasize.