report.Rmd

---
title: "WDV MSA"
author: "Maciej Motyka"
date: "4/8/2020"
output: 
  html_document:
    toc: true
    toc_float: true
    toc_depth: 3
    code_folding: "hide"
    # theme: darkly
    #css: styles.css
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, message = FALSE)
```

# Comparing ClustalW, ClustalO and Muscle alignments
## aBayesQR ml.dist -> NJ trees {.tabset}
Theres a perfect topological agreement between the Muscle, ClustalO, and ClustalW aligned dist.ml distance NJ trees. There are minor branch length differences, mostly in the distance of between WDV-B and WDV-W. ClustalO shows bigger distance than ClustalW and Muscle which show similar distance.

Tree on the left shows true edge lengths. s05 represents WDV-B and is very distant from other sequences which are all WDV-W. Among the WDV-W only s09 forms its own clade which makes sense, as it is the only sample which comes solely from one plant. This clade is the most dissimilar among the WDV-W clades, it includes s09-h3 which is the most distinct haplotype among WDV-W.

Tree on the right doesn't show true distances, but presents the topology much better. s14 and s15 almost form separate clades, in spite of having multiple haplotypes.

### densiTree true distances
```{r fig.height=15, fig.width=12}
drake::loadd("nj_mlDist_ClustO_abqr_001",
             "nj_mlDist_ClustW_abqr_001",
             "nj_mlDist_Muscle_abqr_001")

trees <- c(nj_mlDist_ClustO_abqr_001,
           nj_mlDist_ClustW_abqr_001,
           nj_mlDist_Muscle_abqr_001)

rooted_trees <- lapply(X = trees,
                       FUN = ape::root.phylo,
                       outgroup = "s05_h1_57",
                       resolve.root = T)

class(rooted_trees) <- "multiPhylo"

# Shows true distances between sequences, but makes hard to assess the similarity of topologies
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = F,
                   # jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = T,
                   scale.bar = F
                   )
title("True distances")
```

### densiTree scaled distances
```{r fig.height=20}
# Shows well how similar the topologies are, doesn't show true distances between sequences
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = T,
                   jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = F,
                   label.offset = 0.01,
                   scale.bar = F
                   )
title("Scaled distances")
```

## CliqueSNV t10 tf01 ml.dist -> NJ trees {.tabset}
```{r load and root csnv t10 tf01 trees}
drake::loadd("nj_mlDist_ClustO_csnv_t10_tf01",
             "nj_mlDist_ClustW_csnv_t10_tf01",
             "nj_mlDist_Muscle_csnv_t10_tf01")

trees <- c(nj_mlDist_ClustO_csnv_t10_tf01,
           nj_mlDist_ClustW_csnv_t10_tf01,
           nj_mlDist_Muscle_csnv_t10_tf01)

rooted_trees <- lapply(X = trees,
                       FUN = ape::root.phylo,
                       outgroup = trees[[1]]$tip.label %>% grep("s05_h1", ., value = T),
                       resolve.root = T)

class(rooted_trees) <- "multiPhylo"
```

### Side by side
```{r fig.height=20, fig.width=15}
par(mfrow = c(1,3))
ape::plot.phylo(rooted_trees[[1]],
                main = "ClustalO -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[2]],
                main = "ClustalW -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[3]],
                main = "Muscle -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)
```

### Pairwise ClustO vs ClustW
```{r fig.height=20, fig.keep='all', fig.show='hold'}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[2]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized treesD
```

### Pairwise ClustO vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### Pairwise ClustW vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[2]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### densiTree true distances
```{r fig.height=20, fig.width=12}
# Shows true distances between sequences, but makes hard to assess the similarity of topologies
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = F,
                   # jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = T,
                   scale.bar = F
                   )
title("True distances")
```

### densiTree scaled distances
```{r fig.height=20}
# Shows well how similar the topologies are, doesn't show true distances between sequences
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = T,
                   jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = F,
                   label.offset = 0.01,
                   scale.bar = F
                   )
title("Scaled? distances")
```

## CliqueSNV t5 tf01 ml.dist -> NJ trees {.tabset}
```{r load and root csnv t5 tf01 trees}
drake::loadd("nj_mlDist_ClustO_csnv_t5_tf01",
             "nj_mlDist_ClustW_csnv_t5_tf01",
             "nj_mlDist_Muscle_csnv_t5_tf01")

trees <- c(nj_mlDist_ClustO_csnv_t5_tf01,
           nj_mlDist_ClustW_csnv_t5_tf01,
           nj_mlDist_Muscle_csnv_t5_tf01)

rooted_trees <- lapply(X = trees,
                       FUN = ape::root.phylo,
                       outgroup = trees[[1]]$tip.label %>% grep("s05_h1", ., value = T),
                       resolve.root = T)

class(rooted_trees) <- "multiPhylo"
```

### Side by side
```{r fig.height=20, fig.width=15}
par(mfrow = c(1,3))
ape::plot.phylo(rooted_trees[[1]],
                main = "ClustalO -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[2]],
                main = "ClustalW -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[3]],
                main = "Muscle -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)
```

### Pairwise ClustO vs ClustW
```{r fig.height=20, fig.keep='all', fig.show='hold'}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[2]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized treesD
```

### Pairwise ClustO vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### Pairwise ClustW vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[2]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### densiTree true dist
```{r fig.height=20, fig.width=12}
# Shows true distances between sequences, but makes hard to assess the similarity of topologies
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = F,
                   # jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = T,
                   scale.bar = F
                   )
title("True distances")
```

### densiTree scaled
```{r fig.height=20}
# Shows well how similar the topologies are, doesn't show true distances between sequences
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = T,
                   jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = F,
                   label.offset = 0.01,
                   scale.bar = F
                   )
title("Scaled? distances")
```

## CliqueSNV t5 tf001 ml.dist -> NJ trees {.tabset}
```{r load and root csnv t5 tf001 trees}
drake::loadd("nj_mlDist_ClustO_csnv_t5_tf001",
             "nj_mlDist_ClustW_csnv_t5_tf001",
             "nj_mlDist_Muscle_csnv_t5_tf001")

trees <- c(nj_mlDist_ClustO_csnv_t5_tf001,
           nj_mlDist_ClustW_csnv_t5_tf001,
           nj_mlDist_Muscle_csnv_t5_tf001)

rooted_trees <- lapply(X = trees,
                       FUN = ape::root.phylo,
                       outgroup = trees[[1]]$tip.label %>% grep("s05_h1", ., value = T),
                       resolve.root = T)

class(rooted_trees) <- "multiPhylo"
```

### Side by side
```{r fig.height=20, fig.width=15}
par(mfrow = c(1,3))
ape::plot.phylo(rooted_trees[[1]],
                main = "ClustalO -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[2]],
                main = "ClustalW -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)

ape::plot.phylo(rooted_trees[[3]],
                main = "Muscle -> dist.ml -> NJ",
                use.edge.length = F,
                cex = 1.3,
                label.offset = 0.5)
```

### Pairwise ClustO vs ClustW
```{r fig.height=20, fig.keep='all', fig.show='hold'}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[2]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized treesD
```

### Pairwise ClustO vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[1]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### Pairwise ClustW vs Muscle
```{r fig.height=20}
treespace::plotTreeDiff(tr1 = rooted_trees[[2]],
                       tr2 = rooted_trees[[3]],
                       use.edge.length = FALSE,
                       treesFacing = TRUE,
                       colourMethod = "palette",
                       palette = adegenet::funky,
                       tipMatch = FALSE) # defaults to TRUE; disable when using ladderized trees
```

### densiTree true dist
```{r fig.height=20, fig.width=12}
# Shows true distances between sequences, but makes hard to assess the similarity of topologies
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = F,
                   # jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = T,
                   scale.bar = F
                   )
title("True distances")
```

### densiTree scaled
```{r fig.height=20}
# Shows well how similar the topologies are, doesn't show true distances between sequences
phangorn::densiTree(rooted_trees,
                   type = "p",
                   scaleX = T,
                   jitter = list(amount = 0.05, random = FALSE),
                   alpha = 0.5,
                   col = c("red", "green", "blue"),
                   use.edge.length = F,
                   label.offset = 0.01,
                   scale.bar = F
                   )
title("Scaled? distances")
```

## Session info
```{r session_info, include=TRUE, echo=TRUE, results='markup'}
sessioninfo::session_info()
```