Add ord_val arg to tsg_ind

[jennbaker95]

Problem: When plotting individual values using geom_jitter on top of ranked relative abundance geom_col layer using obs_df arg in plot_ra, ordering of the OTUs on the xaxis (determined by ord_val in taxon_sort_gather df prep) is not preserved. Instead, the individual values layer (which is not ranked in any certain order - defaults to alphanumeric ordering) also forces the plot_ra output to revert to alphanumeric & overwrites the order of the geom_col layer.

mock.gg.mean <- dplyr::filter(otu_df, Type == "Mock") %>% 
                taxon_sort_gather(n=20)
mock.gg <- dplyr::filter(otu_df, Type == "Mock") %>% 
           tsg_ind(n=20, sample_col = "Uniq_ID")
plot_ra(df = mock.gg.mean, df_obs = mock.gg, title = "Mean Ranked Relative Abundance for Mock Positive Controls, Both Runs", fill = "OTU", error_bar = TRUE) + theme(axis.text.x = element_text(angle = 45, hjust = 1, vjust=1), legend.position = "top", plot.margin=unit(c(0.5,0.5,0.5,4), "cm"))

Proposed Solution: Add an ord_val argument and option in tsg_ind to prepare two dfs that have matching ordering. I fixed this problem outside of a user-defined function by modifying source code from taxon_sort_gather.

gg.alltissues.ceft.95 <- filter(otu_df, !is.na(Organ) & Exp_Group == "Ceftriaxone + 95%") 

ord_df <- dplyr::filter(gg.alltissues.ceft.95, Type=="Lung")
otu_order <- rank_tax(ord_df)
gg.alltissues.ceft.95.ind <- data.frame(dplyr::select_if(gg.alltissues.ceft.95,!(stringr::str_detect(colnames(otu_df), "Otu"))), gg.alltissues.ceft.95[, otu_order[1:50]]) %>% 
        join_tax(tax_levels = "Phylum", tax_df = otu_good_taxonomy) %>% 
        dplyr::group_by(OTU, Phylum, Type)
gg.alltissues.ceft.95.ind <- tsg_ind(gg.alltissues.ceft.95.ind, sample_col = "Uniq_ID") 

gg.alltissues.ceft.95.ordL <- taxon_sort_gather(gg.alltissues.ceft.95, facet_var="Type", ord_val = "Lung")

plot_ra(gg.alltissues.ordL, fill = "Type", error_bar = T, df_obs = gg.alltissues.ceft.95.ind) + facet_grid(Type~.)