04-G4.rmd

# IGHV1-3 - G4
```{r}
source("functions.R")
```

```{r, echo=FALSE}
load("data.rda")

chain = "IGH"
func <-  data.frame(allele = names(vgerms[[chain]]), functionality = !grepl("(ORF|P)",sapply(seqinr::getAnnot(vgerms_full[[chain]]), function(x) unlist(strsplit(x,"[|]"))[4])), sign = sapply(seqinr::getAnnot(vgerms_full[[chain]]), function(x) unlist(strsplit(x,"[|]"))[4]), stringsAsFactors = F)

mat <- mat_list$IGH$functional$nonsingle$all$`318`

load("data_frac_new2.rda")
data <- setDT(data_frac$IGH$functional$nonsingle$all$`318`$complete$`95`)
data[,v_call:=paste0(v_gene,"*",v_allele)]
load("alleles_dbs.rda")
allele_db <- alleles_dbs$IGH$functional$nonsingle$all$`318`$complete$`95`
allele_db <- allele_db %>% dplyr::rowwise() %>% dplyr::mutate(gene = alakazam::getGene(or_allele, strip_d = F, omit_nl = F), group = strsplit(gsub(gene, "", new_allele),"[*]")[[1]][1], gene_group = alakazam::getGene(new_allele, strip_d = F, omit_nl = F))
load("functional_groups.rda")
func_groups <- functional_groups$IGH$functional$nonsingle$all$`318`$complete$`95`

cols <- c("#FAAB18", "#1380A1","#990000", "#588300")

pal <- cols %>% 
  newpal(names = c("orangy", "bluish", "redish", "greeny"))

edit_links <- readLines("edit_links.txt")
share_links <- readLines("share_links.txt")
```

```{r,echo=FALSE}
g_group = 'IGHV1-3G4'
group = names(func_groups)[func_groups==g_group]
gr <- allele_db %>% filter(gene_group == g_group) %>% pull(group) %>% unique()
g <- allele_db %>% filter(gene_group == g_group) %>% pull(gene) %>% unique()
```

## Allele appearnce 

The group of `r group` includes `r length(grep(g_group,allele_db$new_allele,value=T))` alleles, `r sum(func$functionality[func$allele %in% allele_db$or_allele[grep(g_group,allele_db$new_allele)]])` out of the alleles are functional.

For each allele we counted the number of appearances across the population, any appearance was considered valid. 

```{r}
allele_appearance(data, g_group, allele_db)
```

## Group alignment

Based on the viewd alleles, we calculated the distance between the germline sequences.

```{r}
v_calls <- unique(data[grepl(g_group,v_gene),v_call])
seq_align(v_calls, allele_db, vgerms, chain, mat, g_group)
```
```{r, fig.width=18,fig.height=30}
v_calls <- unique(data[grepl(g_group,v_gene),v_call])
seq_align2(v_calls, allele_db, vgerms, chain, mat, g_group)
```

## Sequence depth

To examine the potential cutoff we observed the sequence depth for each allele

```{r}
tagList(sequence_depth(data, g_group, allele_db))
```

## Absolute cutoff

We set an initial cutoff of $0.01\%$ to determine the potential genotype priors. For this cutoff we examined the zygousity state, such as homozygousity, heterozygousity and so on.


```{r}

tmp_allele_db <-
  allele_db %>% dplyr::filter(grepl(as.character(g_group), new_allele)) %>%
  dplyr::group_by(new_allele) %>% dplyr::summarise(or_allele = paste0(or_allele, collapse = "/"))

or_allele <-
  setNames(gsub(chain, "", as.character(tmp_allele_db$or_allele)), as.character(gsub(
    paste0(g_group, "[*]"),
    "",
    tmp_allele_db$new_allele
  )))

allele_thresh = 0.5

tmp <- data_cutoff(data, func_g_groups, g_group, allele_thresh, or_allele)

```

With the selected cutoff we saw that there are `r length(unique(tmp %>% filter(is.na(j_call)) %>% arrange(zygousity_state) %>% pull(zygousity_state)))` zygousity states.

```{r eval=knitr::is_html_output(excludes = 'epub'), results = 'asis', echo = F}
source_haplo_usage(g_group, allele_thresh)
```

## Observations

This section is editable by clicking on the edit button below. To refresh the section click on the refresh button

You can access the file also from [here](`r edit_links[grep(g_group,edit_links)]`){target="_blank"}

```{r, message=F, echo=F, warning=F}
knitr::include_url(paste0("https://peresay.shinyapps.io/G2_group/?group=",g_group))
```

## Conclusions

From the results we believe that the cutoff for this group should be 
```{r} 
DT::datatable(data.frame(thresholds = absolute_thresholds_dict[[g_group]]))
```

and for the adjusted states the allele combinations and the relations are stated in the table below.

### Allele specific cutoff

```{r}

tableData <- data_cutoff(data, func_groups, g_group, 5, or_allele) %>% dplyr::group_by(zygousity_state, v_allele) %>% dplyr::summarise(mean_freq = paste0(round(quantile(freq2, 3/4),3),":", round(quantile(freq2, 1/4),3)), v_alleles_abc = unique(v_allele_axis)) %>% dplyr::group_by(zygousity_state, v_alleles_abc) %>% dplyr::summarise(fractions = paste0(mean_freq, collapse = ";"))

DT::datatable(
        tableData,
        options = list(dom = "tipr"),
        selection = 'none',
        colnames = c(
          "Zygousity state" = "zygousity_state",
          "V\nallele" = "v_alleles_abc",
          "IQR range" = "fractions"
        )
      )


# column names change to Zygousity state, alleles combinations, IQR 

```
```{r, out.height= "100%", out.width="100%"}
m <- list(
    l = 50,
    r = 50,
    b = 100,
    t = 100,
    pad = 0.5
)
heatmap_alleles(data, g_group, allele_db, func)%>%
  layout(autosize = F, width = 800, height = 1300)
```