Relative fitness (in NS5A in virus replication)

Growth/Fitness

selection coefficient

unitless

Deep Mutational Scanning , HiSeq 2000 sequencing , PCR amplification

None

None

None

None

None

None

The mutant virus library (12 ml) was used to infect naïve Huh-7.5.1 cells (4million) at M.O.I ∼0.2 with or without BMS-790052 treatment at 20 pM. The supernatant was collected at 144 hpi and used to infect naïve cells for the second round of selection. After two rounds of selection, the viral genome was recovered from the supernatant, and the mutated regions were PCR amplified and processed following the standard sample preparation protocol for HiSeq 2000 sequencing. Each library was tagged with a unique 6-bp molecular barcode sequence, which allows for the identification and study of relative fitness levels in each selection pool.

After determining the number of sequence reads (Reads) for each mutant, we then calculated the frequency of each mutant from each pool and the fitness score in relation to the WT. Any frequency that is lower than 0.0005 will be considered as noise and discarded, since the mutation frequency of HCV is about 1e−5 to 1e−4 nucleotide substitutions per nucleotide per round of genome replication.

The relative fitness score of a given variant was determined as the antilogarithm of the slope of the regression: ln(f_{v,N}/f_{wt,N}) = ln(f_{v,0}/f_{wt,0}) + NlnW

The relative fitness score of each variant in drug treatment was calculated in the same way, but only with 2 rounds of selection in 20 pM drug treatment (Daclatasvir). We then calculated the selection coefficient s:
s = W - 1