Basic plotting of transrate results for varying read depth Trinity mouse assemblies

plots.Rmd

Transrate assembly and contig plots.
========================================================

Assemblies of mouse with 10, 20, 50 and 100 million reads.

# Assembly-level stats

## Load data

```{r}
setwd('~/Dropbox/ongoing_projects/feeding_transcriptome/final')
astats <- read.csv('all_10M_assemblies.csv')
astats$num_pairs <- 10e6
astats$percent_mapping <- with(astats, total_mappings / num_pairs * 100)
astats$pc_good_mapping <- with(astats, good_mappings / num_pairs * 100)
astats$assembly <- gsub(astats$assembly, pattern="\\.corr\\.Trinity\\.fasta", replacement="")
library(reshape2)
astats <- melt(astats, id='assembly')
astats <- astats[complete.cases(astats),]
astats$assembly <- factor(x=astats$assembly, levels=c("10M", "20M", "50M", "100M"), ordered=T)
```

## Plot the metrics

```{r fig.width=7, fig.height=40}
library(ggplot2)
ggplot(astats, aes(x=assembly, y=value, fill=assembly)) + 
  geom_bar(position='dodge', stat='identity') +
  facet_grid(variable ~ ., scales="free") +
  theme(strip.text.y = element_text(angle=0))
```

# Contig-level stats

## Load data

```{r}
library(reshape2)
setwd('~/Dropbox/ongoing_projects/feeding_transcriptome/final')
cstats <- read.csv('all_10M_10M.corr.Trinity.fasta_contigs.csv')
cstats$millionreads <- 10
cstats20 <- read.csv('all_10M_20M.corr.Trinity.fasta_contigs.csv')
cstats20$millionreads <- 20
cstats <- rbind(cstats, cstats20)
cstats50 <- read.csv('all_10M_50M.corr.Trinity.fasta_contigs.csv')
cstats50$millionreads <- 50
cstats <- rbind(cstats, cstats50)
cstats100 <- read.csv('all_10M_100M.corr.Trinity.fasta_contigs.csv')
cstats100$millionreads <- 100
cstats <- rbind(cstats, cstats100)
```

Plot the metrics

## Contig stats

### Length

```{r fig.width=12, fig.height=8}
library(ggplot2)
ggplot(cstats, aes(x=length, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### GC

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=prop_gc, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### GC skew

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=gc_skew, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### AT skew

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=at_skew, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### CpG count

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=cpg_count, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### Cpg ratio

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=cpg_ratio, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### Orf length

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=orf_length, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
```

### Linguistic complexity (k=6)

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=linguistic_complexity_6, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

## Read metrics

### Uncovered bases

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=uncovered_bases, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
```

### Mean coverage

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=mean_coverage, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
ggplot(cstats, aes(y=mean_coverage, x=factor(millionreads))) + 
  geom_violin() +
  labs(colour='Reads (x10^6)') + scale_y_log10()
```

### In bridges

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=in_bridges, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
```

### In bridges

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=in_bridges, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
```

### Edit distance per base

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=edit_distance_per_base, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```

### Low uniqueness bases

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=low_uniqueness_bases, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)') + scale_x_log10()
```

### Low uniqueness bases (proportion)

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=p_low_uniqueness_bases, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```


## Reference-based metrics

### Reference coverage

```{r fig.width=12, fig.height=8}
ggplot(cstats, aes(x=reference_coverage, colour=factor(millionreads))) + 
  geom_density() +
  labs(colour='Reads (x10^6)')
```