Method for detecting a minority genotype

1. A method for detecting a minority genotype of a target nucleic acid of a virus in a population of patients undergoing a population treatment regime, the treatment regime comprising receiving a drug for treating an infection by the virus, the patients being monitored to provide resistance data of the drug, the method comprising:

(a) deep sequencing at least a portion of the target nucleic acid, wherein the deep sequencing is selected from the group consisting of: single molecule real time sequencing, ion semi-conductor sequencing, and bridge polymerase chain reaction sequencing,

(b) detecting, using the deep sequencing results of (a), a presence of variant nucleobases at one or more nucleotide reference positions within the target nucleic acid;

(c) calculating p-values of the variants detected in step (b) and selecting statistically significant variants with the lowest p-values or for which the p-values are below a threshold;

(d) clustering, using non-negative matrix factorization, the statistically significant variants of step (c) into clusters;

(e) correlating the clusters of statistically significant variants with the drug resistance data;

(f) determining a patient treatment regime, different than the population treatment regime, for a patient infected with the virus and having received the drug, from the population of patients or otherwise, based on a determined presence in the patient of a cluster of statistically significant variants correlated with drug resistance in the population as determined in step (e), wherein the patient treatment regime is tailored to the drug resistance; and

(g) administering the determined patient treatment regime to the patient; wherein the virus is hepatitis C and the patient treatment regime and the population treatment regime is each from a class of drugs selected from NS3/4A protease inhibitors, non-nucleoside inhibitors of RdRp and NS5A inhibitors, the selected classes being different for the patient treatment regime and the population treatment regime.

2. The method of claim 1 , wherein said variant detection and variant significance is based on a log likelihood-ratio model for nucleotide variant detection, or a Poisson distribution model for codon variant detection.

3. The method of claim 1 , wherein said variant detection and variant significance is based on a Bayesian model.

4. The method of claim 1 , wherein the statistically significant variants determined in step (c) are subjected to smoothing or a derivative thereof before proceeding to step (d).

5. The method of claim 1 , wherein the statistically significant variants determined in step (c) are subjected to correction by a false discovery rate minimization method before proceeding to step (d).

6. The method of claim 5 , wherein the false discovery rate minimization method includes Bonferroni correction or a derivative thereof.

7. The method of claim 5 , wherein the false discovery rate minimization method includes Benjamini and Hochberg correction or a derivative thereof.

8. The method of claim 1 , further comprising amplifying one or more regions of the target nucleic acid prior to step (a), and wherein said deep sequencing comprises sequencing the one or more amplified regions.

9. The method of claim 1 , wherein the target nucleic acid comprises at least one of the NS3, NS4, and NS5 regions from HCV.

10. The method of claim 1 ,

wherein at least two samples are collected from each patient in the population, wherein each sample of the at least two samples corresponds to a different time point during a course of the population treatment regime, wherein steps (a)-(d) are performed with respect to the target nucleic acid from each of the at least two samples, and

(i) wherein the variant detection results of step (b) include a determination of the frequency for each variant and/or the variant significance results of step (c) include a significance value assigned to each variant, and wherein the method further comprises

(h) for each of one or more variants within the common set of significant variants determined in step (d), comparing the variant frequencies and/or variant significance values for the samples corresponding to different time points to determine whether the variant is a differential variant exhibiting different frequencies and/or significance values during the course of the treatment regime; or

(ii) wherein a plurality of minority genotypes are detected within the samples corresponding to different time points, wherein the variant detection results of step (b) include a determination of the frequency for each variant and/or the variant significance results of step (c) include a significance value assigned to each variant, and wherein the method further comprises

(i) optionally selecting a subset of the plurality of minority genotypes; and

(j) for each of one or more variants within the common set of significant variants determined in step (d) for each minority genotype within said plurality or optional subset thereof, comparing the variant frequencies and/or variant significance values for the samples corresponding to different time points to determine whether the variant is a differential variant exhibiting different frequencies and/or significance values during the course of the treatment regime.

11. The method of claim 10 , wherein a plurality of differential variants are identified in step (j), and wherein the method further comprises

(k) selecting at least a subset of the plurality of differential variants; and

(l) performing two-way hierarchical clustering on (1) the frequencies and/or significance values of the differential variants within said plurality or subset thereof and (2) the samples containing the target nucleic acid in which said differential variants were detected, each sample corresponding to a time point.

12. The method of claim 10 , further comprising confirming at least one differential variant by Sanger sequencing.