speed_comparisons.Rmd

# Speed comparisons between `TreeSE` and `phyloseq`

## Data set inspection

Conversion from `tse` to `pseq` objects employed the following command:

```{r}
# conversion
pseq <- makePhyloseqFromTreeSummarizedExperiment(tse)
```

### mia data sets

```{r}
# load from mia
library(mia)
data_sets <- c("GlobalPatterns", "soilrep", "esophagus", "enterotype")
mapply(data, list = data_sets, package = "mia")
tse <- GlobalPatterns
```

```{r}
# load from phyloseq
library(phyloseq)
data_sets <- c("GlobalPatterns", "soilrep", "esophagus", "enterotype")
mapply(data, list = data_sets, package = "phyloseq")
pseq <- GlobalPatterns
```

### microbiome data sets

```{r}
# load from microbiomeDataSets
library(microbiomeDataSets)
availableDataSets()
tse <- HintikkaXOData()
```

### curatedMetagenomicData data sets

```{r}
# load from curatedMetagenomicData
library(curatedMetagenomicData)
tse <- curatedMetagenomicData("AsnicarF_2021", dryrun = FALSE, counts = TRUE)
tse <- tse[[1]]
```

## Speed comparison of melting

```{r}
# melt tse
start.time1 <- Sys.time()
molten_tse <- mia::meltAssay(tse,
                       abund_values = "counts",
                       add_row_data = TRUE,
                       add_col_data = TRUE)
end.time1 <- Sys.time()
diff_tse <- end.time1 - start.time1
diff_tse
```

```{r}
# melt pseq
start.time2 <- Sys.time()
molten_pseq <- phyloseq::psmelt(pseq)
end.time2 <- Sys.time()
diff_pseq <- end.time2 - start.time2
diff_pseq
```

```{r}
# absolute difference
abs_diff <- diff_pseq - diff_tse
abs_diff
```

```{r}
# sanity check
colnames(molten_tse)
colnames(molten_pseq)
dim(molten_tse)
dim(molten_pseq)
```

## Speed comparison of transformation

```{r}
# transform tse
start.time1 <- Sys.time()
trans_tse <- mia::transformSamples(tse,
                                  method = "log10",
                                  pseudocount = 1)
end.time1 <- Sys.time()
diff_tse <- end.time1 - start.time1
diff_tse
```

```{r}
# transform pseq
start.time2 <- Sys.time()
trans_pseq <- microbiome::transform(pseq,
                                    transform = "log10p",
                                    target = "sample")
end.time2 <- Sys.time()
diff_pseq <- end.time2 - start.time2
diff_pseq
```

```{r}
# absolute difference
abs_diff <- diff_pseq - diff_tse
abs_diff
```

```{r}
# sanity check
# sanity check here
```

## Speed comparison of agglomeration

```{r}
# agglomerate tse
start.time1 <- Sys.time()
tse_phylum <- agglomerateByRank(tse,
                               rank = "Phylum",
                               na.rm = TRUE)
end.time1 <- Sys.time()
diff_tse <- end.time1 - start.time1
diff_tse
```

```{r}
# agglomerate pseq
start.time2 <- Sys.time()
pseq_phylum <- phyloseq::tax_glom(pseq,
                                  taxrank = "Phylum")
# na.rm = TRUE by default in tax_glom
end.time2 <- Sys.time()
diff_pseq <- end.time2 - start.time2
diff_pseq
```

```{r}
# absolute difference
abs_diff <- diff_pseq - diff_tse
abs_diff
```

```{r}
# sanity chek
# sanity check here
colnames(tax_table(pseq))
colnames(tax_table(pseq_phylum))
```

## Speed comparison of alpha diversity estimation

```{r}
# estimate alpha diversity in tse
start.time1 <- Sys.time()
alpha_tse <- mia::estimateDiversity(tse, 
                             abund_values = "counts",
                             index = "shannon", 
                             name = "shannon")
end.time1 <- Sys.time()
diff_tse <- end.time1 - start.time1
diff_tse
```

```{r}
# estimate alpha diversity in pseq
start.time2 <- Sys.time()
alpha_pseq <- microbiome::diversity(pseq,
                     index = "shannon")
end.time2 <- Sys.time()
diff_pseq <- end.time2 - start.time2
diff_pseq
```

```{r}
# sanity check
head(colData(alpha_tse)$shannon)
head(alpha_pseq$shannon)
```