Projecting genes and KEGG Ortholgs on a pathway

[Rahul1711arora]

Hi,

I have found this to be very interesting and played with some data. I'm wondering if I have a list of genes from DEA analysis and the KEGG Ortholog (KO) IDs from another analysis. How can I make sure that for a given pathway, I can project both and highlight which gene is from analysis A and KO is from analysis B but they work together in this pathway. Like glycolysis, how can I project that KO1, KO2, KO3 is from analysis B and the gene1, gene2, and gene3 is from analysis A are both working together in this glycolysis pathway and also label them?

Thanks in advance.

Best,
Rahul

[noriakis]

Hi,

Thank you very much for your interest in ggkegg. I apologize for the long delay in response.

Your use case is very important highlighting KO and gene simultaneously.

I think the projection can be achieved by merging the name of the KO-based pathway into the pathway of species of interest. For HSA, I briefly make an example code highlighting gene changes in red, and KO changes in blue.

library(ggkegg)
g <- pathway("hsa00010")
g2 <- pathway("ko00010")

## add KO names in the HSA pathway
g <- g %N>% mutate(name2=g2 %N>% pull(name))


## Random values
genes <- g %N>% filter(type=="gene") %>% pull(graphics_name) %>% strsplit(", ") %>% sapply("[", 1) %>% unlist() %>% unique()
kos <- g2 %N>% filter(type=="ortholog") %>% pull(name) %>% strsplit(" ") %>% unlist() %>% unique()
group1 <- sample(genes, 10)
group2 <- sample(kos, 10)
val1 <- runif(10) %>% setNames(group1)
val2 <- runif(10) %>% setNames(group2)

val1
#>        HK1       TPI1    ALDH3A1       FBP1     MINPP1        GCK    LDHAL6A 
#> 0.03388509 0.59546988 0.93749792 0.26126452 0.96379131 0.56503199 0.27043752 
#>      ADH1A      PDHA1      ALDH2 
#> 0.91943985 0.02604243 0.71127836
val2
#> ko:K12407 ko:K01834 ko:K03103 ko:K13980 ko:K00170 ko:K01223 ko:K01905 ko:K00172 
#> 0.4896457 0.9828648 0.2256395 0.5059643 0.3850336 0.4675857 0.9891956 0.6774298 
#> ko:K25026 ko:K01622 
#> 0.3978992 0.7771029

g <- g %N>% mutate(val1=node_numeric(val1, name = "graphics_name", sep=", ", remove_dot = TRUE),
                   val2=node_numeric(val2, name = "name2")) %>%
    mutate(width=(xmax-xmin)/2)

library(ggnewscale)
ggraph(g, x=x, y=y) +
    geom_node_rect(aes(fill=val1))+
    scale_fill_distiller(palette = "Reds")+
    new_scale_fill()+
    geom_node_rect(aes(xmin=xmin+width, fill=val2))+
    scale_fill_distiller(palette = "Blues")+
    theme_void()

^{Created on 2024-09-20 with reprex v2.1.1}

Hope this helps your analysis.