Methods for identifying treatment targets based on multiomics data

1. A method for identifying treatment targets for a condition, the method comprising:

receiving a set of multiomics data, the multiomics data including transcriptomics data including data related to the condition;

filtering the transcriptomics data to determine a set of highly expressed genes related to the condition;

determining from the set of highly expressed genes a set of biclusters, each bicluster representing a conditionally co-regulated module of genes; and

determining from the set of biclusters a set of disease-relevant biclusters,

wherein determining a set of biclusters comprises executing a biclustering algorithm using as training data one or more received sets of miRNA targets and/or one or more sets of transcription factor targets, wherein the biclustering algorithm uses as training data a set of transcription factor targets, and wherein the set of transcription factor targets is created by:

extracting from a human genome sequence a set of promoter sequences;

searching the set of promoter sequences for instances of DNA recognition motifs to create a set of instances of DNA recognition motifs; and

identifying in the set of instances of DNA recognition motifs those instances that intersect with digital genomic footprints to create a transcription factor target gene database.

2. A method for identifying treatment targets for a condition, the method comprising:

receiving a set of multiomics data, the multiomics data including transcriptomics data including data related to the condition;

filtering the transcriptomics data to determine a set of highly expressed genes related to the condition;

determining from the set of highly expressed genes a set of biclusters, each bicluster representing a conditionally co-regulated module of genes; and

determining from the set of biclusters a set of disease-relevant biclusters,

wherein determining a set of biclusters comprises executing a biclustering algorithm using as training data one or more received sets of miRNA targets and/or one or more sets of transcription factor targets,

wherein determining from the set of biclusters a set of disease-relevant biclusters comprises:

determining from the set of biclusters a first subset of biclusters, each bicluster in the first subset of biclusters having conditional up/down regulation associated with patient survival in a set of validation data;

determining from the set of biclusters a second subset of biclusters, each bicluster in the second subset biclusters having conditional up/down regulation associated with patient survival or a disease hallmark in the set of multiomics data; and

selecting, as the set of disease-relevant biclusters, biclusters that are in both the first subset of biclusters and the second set of biclusters.

3. A method according to claim 2 , wherein each of the validation data and the multiomics data comprises a set of survival data and a set of transcriptomics data.

4. A method for identifying treatment targets for a condition, the method comprising:

receiving a set of multiomics data, the multiomics data including transcriptomics data including data related to the condition;

filtering the transcriptomics data to determine a set of highly expressed genes related to the condition;

determining from the set of highly expressed genes a set of biclusters, each bicluster representing a conditionally co-regulated module of genes;

determining from the set of biclusters a set of disease-relevant biclusters, wherein determining a set of biclusters comprises executing a biclustering algorithm using as training data one or more received sets of miRNA targets and/or one or more sets of transcription factor targets;

receiving in the multiomics data a set of genomics data related to the condition;

filtering the set of genomics data to determine a set of somatically mutated genes associated with the condition; and

filtering the set of genomics data to determine a set of pathways aggregating somatically mutated genes.

5. A method according to claim 4 , further comprising:

determining a set of bicluster eigengenes from the set of disease-relevant biclusters; and

determining from at least the set of bicluster eigengenes a set of causal transcription factors and a set of causal miRNAs.

6. A method according to claim 5 , wherein determining from at least the set of bicluster eigengenes a set of causal transcription factors and a set of causal miRNAs comprises:

inputting into a network edge orienting algorithm:

the set of bicluster eigengenes;

the set of somatically mutated genes associated with the condition;

the set of pathways aggregating the somatically mutated genes;

a set of miRNAs from the multiomics data; and

a set of transcription factors from the multiomics data.

7. A method for identifying treatment targets for a condition, the method comprising:

receiving a set of multiomics data, the multiomics data including transcriptomics data including data related to the condition;

filtering the transcriptomics data to determine a set of highly expressed genes related to the condition;

determining from the set of highly expressed genes a set of biclusters, each bicluster representing a conditionally co-regulated module of genes; and

determining from the set of biclusters a set of disease-relevant biclusters, wherein determining a set of biclusters comprises executing a biclustering algorithm using as training data one or more received sets of miRNA targets and/or one or more sets of transcription factor targets;

expanding the set of mechanistic transcription factors to include other transcription factors in a same family as each of the set of mechanistic transcription factors;

finding a set of correlated transcription factors in the expanded set of mechanistic transcription factors that are correlated with bicluster eigengenes;

determining a first set of transcription factors that have both causal and mechanistic support for regulation of the same bicluster, by taking the intersection of the set of correlated transcription factors and the set of causal transcription factors;

determining a second set of transcription factors that have both causal and mechanistic support for regulation of the same bicluster, by inputting the set of causal transcription factors into an analysis of motif enrichment algorithm; and

taking the union of the first set of transcription factors and the second set of transcription factors to produce a set of treatment targets including causal and mechanistic transcription factors,

wherein determining from the set of highly expressed genes a set of biclusters further comprises:

determining a set of mechanistic transcription factors; and

determining a set of mechanistic miRNAs.

8. A method according to claim 7 , further comprising:

evaluating, for treatment targets in the set of treatment targets, whether the treatment target is positively or negatively associated with survival;

determining the regulator function of the treatment target; and

determining whether to decrease expression or activity (knock down) or increase expression or activity of the treatment target to achieve a therapeutic effect for the condition.

9. A method of selecting a combination therapy to inhibit growth of neoplastic cells in a mammalian subject, the method comprising:

identifying two or more treatment targets, wherein the two or more treatment targets are independently selected from the group consisting of transcription factors and miRNAs, and determining whether increased expression/activity or decreased expression/activity of the two or more treatment targets is expected to decrease growth of the neoplastic cells, according to claim 8 ; and

selecting as a combination therapy two or more agents to modulate the treatment targets in the directions expected to decrease growth of the neoplastic cells.

10. The method according to claim 9 , that comprises determining that a decreased expression or activity of two or more targets is expected to decrease growth of the neoplastic cells, and that comprises selecting as the combination therapy two or more interfering RNAs to decrease expression of the two or more targets.

11. A method of treatment of a mammalian subject to inhibit growth of neoplastic cells, the method comprising:

identifying two or more treatment targets, wherein the two or more treatment targets are independently selected from the group consisting of transcription factors and miRNAs, and determining whether increased expression/activity or decreased expression/activity of the two or more treatment targets is expected to decrease growth of the neoplastic cells, according to claim 8 ; and

administering agents to the mammalian subject in amounts effective to modulate the treatment targets in the directions expected to decrease growth of the neoplastic cells.

12. The method according to claim 11 that comprises determining that decreased expression/activity of two or more treatment targets is expected to decrease growth of the neoplastic cells, and the administering step comprises administering interfering RNA molecules selected for the two or more treatment targets, to decrease expression/activity of the two or more targets.

13. A method for identifying treatment targets for a condition, the method comprising:

receiving a set of multiomics data, the multiomics data including transcriptomics data including data related to the condition;

filtering the transcriptomics data to determine a set of highly expressed genes related to the condition;

determining from the set of highly expressed genes a set of biclusters, each bicluster representing a conditionally co-regulated module of genes; and

determining from the set of biclusters a set of disease-relevant biclusters, wherein determining a set of biclusters comprises executing a biclustering algorithm using as training data one or more received sets of miRNA targets and/or one or more sets of transcription factor targets;

determining restricted set of mechanistic miRNAs by restricting the set of mechanistic miRNAs to include only miRNAs that exhibit anti-correlated expression with bicluster eigengenes; and

taking the union of the restricted set of mechanistic miRNAs and the set of causal miRNAs to produce a set of treatment targets including causal and mechnanistic miRNAs, wherein determining from the set of highly expressed genes a set of biclusters further comprises:

determining a set of mechanistic transcription factors; and

determining a set of mechanistic miRNAs.