octopus.glm.nb: Fit a Negative Binomial Generalized Linear Model
In WeiZhang317/octopus: Tools for GLM NB analysis

Description Usage Arguments Value Examples

Fit glm.nb , add anova to result, add lsm to result.

Before fitting, data and factors will be merged by row.names, mismatching rows will be removed.

If formula start with ~, will fit model with each data column of data as respond y.

1	octopus.glm.nb(formula, data, factors, specs4lsmeans = NULL)

`formula`	formula
`data`	normalized reads, expecting gene as column, sample as row
`factors`	factors, expecting factor name as column, sample as row
`specs4lsmeans`	specs for lsmeans

model if formula has respond y; flattened anova table, lsm table and se table if formula starts with ~.

{
    data(sample_keys) # facotrs
    for(cur_col in colnames(sample_keys)){ sample_keys[,cur_col] <- as.factor(sample_keys[,cur_col]) }

    data(sample_reads) # reads
    # reads normalization
    norm.reads <- octopus.normalize(sample_reads)

    print(row.names(sample_keys))
    print(row.names(norm.reads))
    # transpose it to meet input requirements of founction octopus.glm.nb()
    norm.reads <- data.frame(t(norm.reads))
    print(row.names(norm.reads))

    # fit model with first gene column, column name is Bcin01g00040.1
    result <- octopus.glm.nb(Bcin01g00040.1 ~ Experiment + Experiment/GrowingFlat + Experiment/GrowingFlat/AgarFlat + Isolate + HostGenotype + HostGenotype*Isolate
                             ,norm.reads
                             ,sample_keys)
    print(result$anova)

    # calculate LSMean
    result$lsmeans <- lsmeans(result,~ Isolate | HostGenotype)
    summary(result$lsmeans)

    # fit model and pull LSMean in one call
    result <- octopus.glm.nb(Bcin01g00040.1 ~ Experiment + Experiment/GrowingFlat + Experiment/GrowingFlat/AgarFlat + Isolate + HostGenotype + HostGenotype*Isolate
                             ,norm.reads
                             ,sample_keys
                             ,specs4lsmeans = ~ Isolate | HostGenotype)
    print(result$anova)
    summary(result$lsmeans)

    # iterate through all data
    result <- octopus.glm.nb( ~ Experiment + Experiment/GrowingFlat + Experiment/GrowingFlat/AgarFlat + Isolate + HostGenotype + HostGenotype*Isolate
                              ,norm.reads
                              ,sample_keys
                              ,specs4lsmeans = ~ Isolate | HostGenotype)
    print(result$anova)
    print(result$lsm)
    print(result$se)

  }