scp.Rmd

---
title: "Single Cell Proteomics data processing and analysis"
author:
    - name: Laurent Gatto
    - name: Christophe Vanderaa
output:
    BiocStyle::html_document:
        self_contained: yes
        toc: true
        toc_float: true
        toc_depth: 2
        code_folding: show
bibliography: scp.bib
date: "`r BiocStyle::doc_date()`"
package: "`r BiocStyle::pkg_ver('scp')`"
vignette: >
    %\VignetteIndexEntry{Single Cell Proteomics data processing and analysis}
    %\VignetteEngine{knitr::rmarkdown}
    %\VignetteEncoding{UTF-8}
---

```{r setup, include = FALSE}
knitr::opts_chunk$set(
    collapse = TRUE,
    comment = "#>",
    crop = NULL
    ## cf https://stat.ethz.ch/pipermail/bioc-devel/2020-April/016656.html
)
```

# The `scp` package

The `scp` package is used to process and analyse mass spectrometry
(MS)-based single cell proteomics (SCP) data. The functions rely on a
specific data structure that wraps
[`QFeatures`](https://rformassspectrometry.github.io/QFeatures/)
objects (@Gatto2023-ry) around
[`SingleCellExperiment`](http://bioconductor.org/packages/release/bioc/html/SingleCellExperiment.html)
objects (@Amezquita2019-bf). This data structure could be seen as
Matryoshka dolls were the `SingleCellExperiment` objects are small
dolls contained in the bigger `QFeatures` doll.

The `SingleCellExperiment` class provides a dedicated framework for
single-cell data. The `SingleCellExperiment` serves as an interface to
many cutting-edge methods for processing, visualizing and analysis
single-cell data. More information about the `SingleCellExperiment`
class and associated methods can be found in the
[OSCA book](http://bioconductor.org/books/release/OSCA/).

The `QFeatures` class is a data framework dedicated to manipulate and
process MS-based quantitative data. It preserves the relationship
between the different levels of information: peptide to spectrum match
(PSM) data, peptide data and protein data. The `QFeatures` package
also provides an interface to many utility functions to streamline the
processing MS data. More information about MS data analysis tools can
be found in the
[RforMassSpectrometry project](https://www.rformassspectrometry.org/).

```{r scp_framework, results='markup', fig.cap="`scp` relies on `SingleCellExperiment` and `QFeatures` objects.", echo=FALSE, out.width='100%', fig.align='center'}
knitr::include_graphics("./figures/SCP_framework.png", error = FALSE)
```

Before running the vignette we need to load the `scp` package.

```{r load_scp, message = FALSE}
library("scp")
```

We also load `ggplot2` and `dplyr` for convenient data
manipulation and plotting.

```{r load_other_package, message = FALSE}
library("ggplot2")
library("dplyr")
```

# Before you start

This vignette will guide you through some common steps of mass
spectrometry-based single-cell proteomics (SCP) data analysis. SCP is
an emerging field and further research is required to develop a
principled analysis workflow. Therefore, we **do not guarantee** that the
steps presented here are the best steps for this type of data analysis.
This vignette performs the steps that were described in the SCoPE2
landmark paper (@Specht2021-jm) and that were reproduced in another
work using the `scp` package (@Vanderaa2021-ue). The replication on
the full SCoPE2 dataset using `scp` is available in
[this vignette](https://uclouvain-cbio.github.io/SCP.replication/articles/SCoPE2.html).
We hope to convince the reader that,
although the workflow is probably not optimal, `scp` has the full
potential to perform standardized and principled data analysis. All
functions presented here are comprehensively documented, highly modular,
can easily be extended with new algorithms. Suggestions, feature
requests or bug reports are warmly welcome. Feel free to open an issue
in the
[GitHub repository](https://github.com/UCLouvain-CBIO/scp/issues).

This workflow can be applied to any MS-based SCP data. The minimal
requirement to follow this workflow is that the data should contain
the following information:

- `runCol`/`Raw.file`: field in the feature data and the sample data
  that gives the names of MS acquisition runs or files.
- `quantCols`: field in the sample data that links to columns in the
  quantification data and that allows to link samples to MS channels
  (more details in another
  [vignette](https://uclouvain-cbio.github.io/scp/articles/read_scp.html)).
- `SampleType`: field in the sample data that provides the type of
  sample that is acquired (carrier, reference, single-cell,...). Only
  needed for multiplexing experiments.
- `Potential.contaminant`: field in the feature data that marks
  contaminant peptides.
- `Reverse`: field in the feature data that marks reverse peptides.
- `PIF`: field in the feature data that provides spectral purity.
- `PEP` or `dart_PEP`: field in the feature data that provides peptide
  posterior error probabilities.
- `Modified.sequence`: field in the feature data that provides the
  peptide identifiers.
- `Leading.razor.protein`: field in the feature data that provides the
  protein identifiers.
- At least one field in the feature data that contains quantification
  values. In this case, there are 16 quantification columns named as
  `Reporter.intensity.` followed by an index (`1` to `16`).

Each required field will be described more in detail in the
corresponding sections. Names can be adapted by the user to more
meaningful ones or adapted to other output tables.

# Read in SCP data

The first step is to read in the PSM quantification table generated
by, for example, MaxQuant (@Tyanova2016-yj). We created a small
example data by subsetting the MaxQuant `evidence.txt` table provided
in the SCoPE2 landmark paper (@Specht2021-jm). The `mqScpData` table
is a typical example of what you would get after reading in a CSV file
using `read.csv` or `read.table`. See `?mqScpData` for more
information about the table content.

```{r}
data("mqScpData")
```

We also provide an example of a sample annotation table that provides
useful information about the samples that are present in the example
data. See `?sampleAnnotation` for more information about the table
content.

```{r}
data("sampleAnnotation")
```

As a note, the example sample data contains 5 different types of samples
(`SampleType`) that can be found in a TMT-based SCP data set:

```{r}
table(sampleAnnotation$SampleType)
```

* The carrier channels (`Carrier`) contain 200 cell equivalents and
  are meant to boost the peptide identification rate.
* The normalization channels (`Reference`) contain 5 cell equivalents
  and are used to partially correct for between-run variation.
* The unused channels (`Unused`) are channels that are left empty due
  to isotopic cross-contamination by the carrier channel.
* The negative control channels (`Blank`) contain samples that do not
  contain any cell but are processed as single-cell samples.
* The single-cell sample channels contain the single-cell samples of
  interest, that are macrophage (`Macrophage`) or monocyte
  (`Monocyte`).

Using `readSCP`, we combine both tables in a `QFeatures` object
formatted as described above.

```{r readSCP}
scp <- readSCP(assayData = mqScpData,
               colData = sampleAnnotation,
               runCol = "Raw.file",
               removeEmptyCols = TRUE)
scp
```

See here that the 3 first assays contain 11 columns that correspond to
the TMT-11 labels and the last assay contains 16 columns that
correspond to the TMT-16 labels.

**Important**: More details about the usage of `readSCP()` and how to
read your own data set are provided in the `Load data using readSCP`
[vignette](https://uclouvain-cbio.github.io/scp/articles/read_scp.html).

Another way to get an overview of the scp object is to plot the
`QFeatures` object. This will create a graph where each node is an
assay and links between assays are denoted as edges.

```{r plot1}
plot(scp)
```


# Clean missing data

All single-cell data contain many zeros. The zeros can be biological
zeros or technical zeros and differentiating between the two types is
not a trivial task. To avoid artefacts in downstream steps, we replace
the zeros by the missing value `NA`. The `zeroIsNA` function takes the
`QFeatures` object and the name(s) or index/indices of the assay(s) to
clean and automatically replaces any zero in the selected quantitative
data by `NA`.

```{r zeroIsNA}
scp <- zeroIsNA(scp, i = 1:4)
```

# Filter PSMs

A common steps in SCP is to filter out low-confidence PSMs. Each PSM
assay contains feature meta-information that are stored in the
`rowData` of the assays. The `QFeatures` package allows to quickly
filter the rows of an assay by using these information. The available
variables in the `rowData` are listed below for each assay.

```{r rowDataNames}
rowDataNames(scp)
```

## Filter features based on feature annotations

Below are some examples of criteria that are used to identify
low-confidence. The information is readily available since this was
computed by MaxQuant:

- Remove PSMs that are matched to contaminants
- Remove PSMs that are matched to the decoy database
- Keep PSMs that exhibit a high PIF (parental ion fraction),
indicative of the purity of a spectrum

We can perform this filtering using the `filterFeatures` function from
`QFeatures`. `filterFeatures` automatically accesses the feature
annotations and selects the rows that meet the provided condition(s). For
instance, `Reverse != "+"` keeps the rows for which the `Reverse`
variable in the `rowData` is not `"+"` (*i.e.* the PSM is not matched
to the decoy database).

```{r filter_psms}
scp <- filterFeatures(scp,
                      ~ Reverse != "+" &
                          Potential.contaminant != "+" &
                          !is.na(PIF) & PIF > 0.8)
```

## Filter assays based on detected features

To avoid proceeding with failed runs, another interesting filter is to
remove assays with too few features. If a batch contains less than,
for example, 150 features we can then suspect something wrong happened
in that batch and it should be removed. Using `dims`, we can query the
dimensions (hence the number of features and the number of samples) of
all assays contained in the dataset.

```{r dims}
dims(scp)
```

Actually, a `QFeatures` object can be seen as a three-order array:
$features \times samples \times assay$. Hence, `QFeatures` supports
three-order subsetting `x[rows, columns, assays]`. We first select the
assays that have sufficient PSMs (the number of rows is greater than
150), and then subset the `scp` object for the assays that meet the
criterion.

```{r filter_assays}
keepAssay <- dims(scp)[1, ] > 150
scp <- scp[, , keepAssay]
scp
```

Notice the `190321S_LCA10_X_FP97_blank_01` sample was removed because
it did not contain sufficient features, as expected from a blank
run. This could also have been the case for failed runs.

## Filter features based on SCP metrics

Another type of filtering is specific to SCP. In the SCoPE2 analysis,
the authors suggest a filters based on the sample to carrier ratio
(SCR), that is the reporter ion intensity of a single-cell sample
divided by the reporter ion intensity of the carrier channel (200
cells) from the same batch. It is expected that the carrier
intensities are much higher than the single-cell intensities.

The SCR can be computed using the `computeSCR` function from
`scp`. The function must be told which channels are the samples that
must be divided and which channel contains the carrier. This
information is provided in the sample annotations and is accessed using
the `colData`, under the `SampleType` field.

```{r show_annotation}
table(colData(scp)[, "SampleType"])
```

In this dataset, `SampleType` gives the type of sample that is present
in each TMT channel. The SCoPE2 protocole includes 5 types of samples:

* The carrier channels (`Carrier`) contain 200 cell equivalents and
  are meant to boost the peptide identification rate.
* The normalization channels (`Reference`) contain 5 cell equivalents
  and are used to partially correct for between-run variation.
* The unused channels (`Unused`) are channels that are left empty due
  to isotopic cross-contamination by the carrier channel.
* The negative control channels (`Blank`) contain samples that do not
  contain any cell but are processed as single-cell samples.
* The single-cell sample channels contain the single-cell samples of
  interest, that are macrophage (`Macrophage`) or monocyte
  (`Monocyte`).

The `computeSCR` function expects the following input:

* The `QFeatures` dataset
* The assay name(s) or index/indices for which the SCR should be
  computed
* `colvar`: the variable in the sample annotations (`colData`) that
  hold the information used to discriminate sample channels from
  carrier channels.
* `carrierPattern`: a string pattern (following regular expression
  syntax) that identifies the carrier channel in each batch.
* `samplePattern`: a string pattern (following regular expression
  syntax) that identifies the samples to divide.

Optionally, you can also provide the following arguments:

* `rowDataName`: the name of the column in the `rowData` where to
  store the computed SCR for each feature.
* `sampleFUN`: when multiple samples are present in an assay, there
  are as many SCR as there are samples that need to be summarized to a
  single value per feature. `sampleFUN` tells which function to use
  for summarizing the sample values before computing the SCR; the
  default is the `mean`.
* `carrierFUN`: some designs might include several carriers per run
  (not the case in this example). Similarly to `sampleFUN`,
  `carrierFUN` tells which function to use for summarizing the carrier
  values before computing the SCR; the default is the same function as
  `sampleFUN`.

The function creates a new field in the `rowData` of the assays. We
compute the average SCR for each PSM and store it in the corresponding
`rowData`, under the `MeanSCR` column.

```{r computeSCR}
scp <- computeSCR(scp,
                  i = 1:3,
                  colvar = "SampleType",
                  carrierPattern = "Carrier",
                  samplePattern = "Macrophage|Monocyte",
                  sampleFUN = "mean",
                  rowDataName = "MeanSCR")
```

Before applying the filter, we plot the distribution of the average
SCR. We collect the `rowData` from several assays in a single table
`DataFrame` using the `rbindRowData` function from `QFeatures`.

```{r plot_SCR, warning=FALSE, message=FALSE}
rbindRowData(scp, i = 1:3) |>
    data.frame() |>
    ggplot(aes(x = MeanSCR)) +
    geom_histogram() +
    geom_vline(xintercept = c(1/200, 0.1),
               lty = c(2, 1)) +
    scale_x_log10()
```

The expected ratio between single cells and the carrier is 1/200
(dashed line). We can see that the distribution mode is slightly
shifted towards higher ratios with a mode around 0.01. However, there
are a few PSMs that stand out of the distribution and have a much
higher signal than expected, indicating something wrong happened
during the quantification of those PSMs. We therefore filter out PSMs
with an average SCR higher than 0.1 (solide line). This is again
easily performed using the `filterFeatures` functions.

```{r filter_SCR}
scp <- filterFeatures(scp,
                      ~ !is.na(MeanSCR) &
                          MeanSCR < 0.1)
```

## Filter features to control for FDR

Finally, we might also want to control for false discovery rate (FDR).
MaxQuant already computes posterior error probabilities (PEP), but
filtering on PEPs is too conservative (@Kall2008-hb) so we provide the
`pep2qvalue` function to convert PEPs to q-values that are directly
related to FDR. We here compute the q-values from the PEP (`dart_PEP`)
across all 3 assays. `dart_PEP` contains the PEP values that have been
updated using the DART-ID algorithm (@Chen2019-uc). The function will
store the results in the `rowData`, we here asked to name the new
column `qvalue_PSMs`.

```{r compute_qvalues}
scp <- pep2qvalue(scp,
                  i = 1:3,
                  PEP = "dart_PEP",
                  rowDataName = "qvalue_PSMs")
```

We also allow to compute q-values at peptide or protein level rather
than PSM. In this case, you need to supply the `groupBy` argument.
Suppose we want to compute the q-values at protein level, we can fetch
the protein information stored under `Leading.razor.protein` in the
`rowData`. This time, we store the q-values in a new field called
`qvalue_proteins`.

```{r compute_qvalues2}
scp <- pep2qvalue(scp,
                  i = 1:3,
                  PEP = "dart_PEP",
                  groupBy = "Leading.razor.protein",
                  rowDataName = "qvalue_proteins")
```

We can now filter the PSM to control, let's say, the protein FDR at
1\%. This can be performed using `filterFeatures` because the q-values
were stored in the `rowData`.

```{r filter_FDR}
scp <- filterFeatures(scp,
                      ~ qvalue_proteins < 0.01)
```

# Process the PSM data

## Relative reporter ion intensity

In order to partialy correct for between-run variation, SCoPE2
suggests computing relative reporter ion intensities. This means that
intensities measured for single-cells are divided by the reference
channel containing 5-cell equivalents. We use the `divideByReference`
function that divides channels of interest by the reference
channel. Similarly to `computeSCR`, we can point to the samples and
the reference columns in each assay using the annotation contained in
the `colData`.

We here divide all columns (using the regular expression wildcard `.`)
by the reference channel (`Reference`).

```{r divideByReference}
scp <- divideByReference(scp,
                         i = 1:3,
                         colvar = "SampleType",
                         samplePattern = ".",
                         refPattern = "Reference")
```

# Aggregate PSM data to peptide data

Now that the PSM assays are processed, we can aggregate them to
peptides. This is performed using the `aggregateFeaturesOverAssays`
function. For each assay, the function aggregates several PSMs into a
unique peptide. This is best illustrated by the figure below.

```{r features_aggregation, results = 'markup', fig.cap = "Conceptual illustration of feature aggregation.", echo=FALSE, out.width='100%', fig.align='center'}
knitr::include_graphics("./figures/feature_aggregation.png", error = FALSE)
```

Remember there currently are three assays containing the PSM data.

```{r show_agg_psms}
scp
```

The PSMs are aggregated over the `fcol` feature variable, here the
modified peptide sequence. We also need to supply an aggregating
function that will tell how to combine the quantitative data of the
PSMs to aggregate. We here aggregate the PSM data using the median
value per sample thanks to the `matrixStats:colMedians()`
function. Other functions can be used and we refer to
`?aggregateFeatures` for more information about available aggregation
functions. The `aggregateFeaturesOverAssays()` function will create a
new assay for each aggregated assay. We name the aggregated assays
using the original names and appending `peptides_` at the start.

```{r aggregate_peptides, message = FALSE}
scp <- aggregateFeaturesOverAssays(scp,
                                   i = 1:3,
                                   fcol = "Modified.sequence",
                                   name = paste0("peptides_", names(scp)),
                                   fun = matrixStats::colMedians, na.rm = TRUE)
```

Notice that 3 new assays were created in the `scp` object. Those new
assays contain the aggregated features while the three first assays
are unchanged. This allows to keep track of the data processing.

```{r show_agg_peptides}
scp
```

Under the hood, the `QFeatures` architecture preserves the
relationship between the aggregated assays. See `?AssayLinks` for more
information on relationships between assays. This is illustrated on
the `QFeatures` plot:

```{r plot2}
plot(scp)
```


# Join the SCoPE2 sets in one assay

Up to now, we kept the data belonging to each MS run in separate
assays. We now combine all batches into a single assay. This is done
using the `joinAssays` function from the `QFeatures` package. Note
that we now use the aggregated assays, so assay 4 to 6.

```{r joinAssays}
scp <- joinAssays(scp,
                  i = 4:6,
                  name = "peptides")
```

In this case, one new assay is created in the `scp` object that
combines the data from assay 4 to 6. The samples are always distinct
so the number of column in the new assay (here $48$) will always
equals the sum of the columns in the assays to join (here $16 + 16 +
16$). The feature in the joined assay might contain less features than
the sum of the rows of the assays to join since common features
between assays are joined in a single row.

```{r plot3}
plot(scp)
```

# Filter single-cells

Another common step in single-cell data analysis pipelines is to
remove low-quality cells. After subsetting for the samples of interest,
we will use 2 metrics: the median relative intensities per cell and
the median coefficient of variation (CV) per cell.

## Filter samples of interest

We can subset the cells of interest, that is the negative control
samples, the macrophages and the monocytes. This
can easily be done by taking advantage of the `colData` and the
subsetting operators. Recall that `QFeatures` objects support
three-order subsetting, `x[rows, columns, assays]`, where columns are
the samples of interest.

```{r subset samples}
scp
scp <- scp[, scp$SampleType %in% c("Blank", "Macrophage", "Monocyte"), ]
```

The subsetting removes unwanted samples from all assays. The filtered
data set contains the same number of assays with the same number of
features, but the number of columns (hence sampled) decreased.

```{r show_cell_filter}
scp
```


## Filter based on the median relative intensity

We compute the median relative reporter ion intensity for each cell
separately and apply a filter based on this statistic. This procedure
recalls that of library size filtering commonly performed in scRNA-Seq
data analysis, where the library size is the sum of the counts in each
single cell. We compute and store the median intensity in the
`colData`.

```{r compute_colMedians}
medians <- colMedians(assay(scp[["peptides"]]), na.rm = TRUE)
scp$MedianRI <- medians
```

Looking at the distribution of the median per cell can highlight
low-quality cells.

```{r plot_medianRI, warning=FALSE, message=FALSE}
colData(scp) |>
    data.frame() |>
    ggplot() +
    aes(x = MedianRI,
        y = SampleType,
        fill = SampleType) +
    geom_boxplot() +
    scale_x_log10()
```

The negative control samples should not contain any peptide
information and are therefore used to assess the amount of background
signal. The graph above confirms that the signal measured in
single-cells (macrophages and monocytes) is above the background
signal, hence no filtering is needed. Would it not be the case, the
same procedure as in the previous section can be used for selecting
the cells that have an associated median RI lower that a defined
threshold.

## Filter based on the median CV

The median CV measures the consistency of quantification for a group
of peptides that belong to a protein. We remove cells that exhibit
high median CV over the different proteins. We compute the median CV
per cell using the `computeMedianCV` function from the `scp`
package. The function takes the `peptides` assay and computes the CV
for each protein in each cell. To perform this, we must supply the
name of the `rowData` field that contains the protein information
through the `groupBy` argument. We also only want to compute CVs if we
have at least 5 peptides per protein. Finally, we also perform a
normalization and divide the columns by the median. The computed
median CVs are automatically stored in the `colData` under the name
that is supplied, here `MedianCV`.

```{r medianCVperCell}
scp <- medianCVperCell(scp,
                       i = 1:3,
                       groupBy = "Leading.razor.protein",
                       nobs = 5,
                       norm = "div.median",
                       na.rm = TRUE,
                       colDataName = "MedianCV")
```

The computed CVs are stored in the `colData` of the `peptides` assay
and holds the median CV per cell computed using at least 5
observations (peptides). The main interest of computing the median CV
per cell is to filter cells with reliable quantification. The negative
control samples are not expected to have reliable quantifications and
hence can be used to estimate an empirical null distribution of the
CV. This distribution helps defining a threshold that filters out
single-cells that contain noisy quantification.

```{r plot_medianCV, message = FALSE, warning = FALSE}
getWithColData(scp, "peptides") |>
    colData() |>
    data.frame() |>
    ggplot(aes(x = MedianCV,
               fill = SampleType)) +
    geom_boxplot() +
    geom_vline(xintercept = 0.65)
```

We can see that the protein quantification for single-cells are much
more consistent than for negative control samples. Based on the
distribution of the negative controls, we decide to keep the cells
that have a median CV lower than 0.65. Note this example is inaccurate
because the null distribution is based on only 3 negative controls,
but more sets could lead to a better estimation of the CV null
distribution.

```{r create_filter}
scp <- scp[, !is.na(scp$MedianCV) & scp$MedianCV < 0.65, ]
```

We can now remove the negative controls since all QC metrics are now
computed.

```{r remove_NC}
scp <- scp[, scp$SampleType != "Blank", ]
```

# Process the peptide data

In this vignette, the peptide data are further processed before
aggregation to proteins. The steps are: normalization, filter peptides
based on missing data and log-transformation.

## Normalization

The columns (samples) of the peptide data are first normalized by
dividing the relative intensities by the median relative
intensities. Then, the rows (peptides) are normalized by dividing the
relative intensities by the mean relative intensities. The normalized
data is stored in a separate assay.  This normalization procedure is
suggested in the SCoPE2 analysis and is applied using the `sweep`
method. Beside the dataset and the assay to normalize, the method
expects a `MARGIN`, that is either row-wise (`1`) or column-wise (`2`)
transformation, the `FUN` function to apply and `STATS`, a vector of
values to apply. More conventional normalization procedure can be
found in `?QFeatures::normalize`.

```{r normalize_scale}
## Divide columns by median
scp <- sweep(scp,
             i = "peptides",
             MARGIN = 2,
             FUN = "/",
             STATS = colMedians(assay(scp[["peptides"]]), na.rm = TRUE),
             name = "peptides_norm_col")
## Divide rows by mean
scp <- sweep(scp,
             i = "peptides_norm_col",
             MARGIN = 1,
             FUN = "/",
             STATS = rowMeans(assay(scp[["peptides_norm_col"]]),  na.rm = TRUE),
             name = "peptides_norm")
```

Notice each call to `sweep` created a new assay. Let's have a look to
the current stat of the `QFeatures` plot:

```{r plot4}
plot(scp)
```

## Remove peptides with high missing rate

Peptides that contain many missing values are not
informative. Therefore, another common procedure is to remove higly
missing data. In this example, we remove peptides with more than 99 \%
missing data. This is done using the `filterNA` function from
`QFeatures`.

```{r filterNA}
scp <- filterNA(scp,
                i = "peptides_norm",
                pNA = 0.99)
```

## Log-transformation

In this vignette, we perform log2-transformation using the
`logTransform` method from `QFeatures`. Other log-transformation can
be applied by changing the `base` argument.

```{r logTransform}
scp <- logTransform(scp,
                    base = 2,
                    i = "peptides_norm",
                    name = "peptides_log")
```

Similarly to `sweep`, `logTransform` creates a new assay in `scp`.

# Aggregate peptide data to protein data

Similarly to aggregating PSM data to peptide data, we can aggregate
peptide data to protein data using the `aggregateFeatures` function.

```{r aggregate_proteins}
scp <- aggregateFeatures(scp,
                         i = "peptides_log",
                         name = "proteins",
                         fcol = "Leading.razor.protein",
                         fun = matrixStats::colMedians, na.rm = TRUE)
```

The only difference between `aggregateFeatures` and
`aggregateFeaturesOverAssays` is that the second function can
aggregate several assay at once whereas the former only takes one
assay to aggregate. Hence, only a single assay, `proteins`, was
created in the `scp` object.

```{r show_agg_proteins}
scp
```

After the second aggregation, the `proteins` assay in this example
contains quantitative information for 89 proteins in 15 single-cells.

# Process the protein data

The protein data is further processed in three steps: normalization,
imputation (using the KNN algorithm) and batch correction (using the
`ComBat` algorithm).

## Normalization

Normalization is performed similarly to peptide normalization. We use
the same functions, but since the data were log-transformed at the
peptide level, we subtract by the statistic (median or mean) instead
of dividing.

```{r normalize_center}
## Center columns with median
scp <- sweep(scp, i = "proteins",
             MARGIN = 2,
             FUN = "-",
             STATS = colMedians(assay(scp[["proteins"]]),
                                na.rm = TRUE),
             name = "proteins_norm_col")
## Center rows with mean
scp <- sweep(scp, i = "proteins_norm_col",
             MARGIN = 1,
             FUN = "-",
             STATS = rowMeans(assay(scp[["proteins_norm_col"]]),
                              na.rm = TRUE),
             name = "proteins_norm")
```


## Imputation

The protein data contains a lot of missing values.

```{r missingness}
scp[["proteins_norm"]] |>
    assay() |>
    is.na() |>
    mean()
```

The average missingness in the `proteins` assay is around 25
\%. Including more samples and hence more batches can increase the
missingness up to 70 \% as seen for the complete SCoPE2 dataset
(@Specht2021-jm). Whether imputation is beneficial or deleterious for
the data will not be discussed in this vignette.  But taking those
missing value into account is essential to avoid artefacts in
downstream analyses. The data imputation is performed using the K
nearest neighbors algorithm, with k = 3. This is available from the
`impute()` mehtod.  More details about the arguments can be found in
`?impute::impute.knn`.

```{r impute}
scp <- impute(scp,
              i = "proteins_norm",
              name = "proteins_imptd",
              method = "knn",
              k = 3, rowmax = 1, colmax= 1,
              maxp = Inf, rng.seed = 1234)
```

Note that after imputation, no value are missing.

```{r missingness_imputed}
scp[["proteins_imptd"]] |>
    assay() |>
    is.na() |>
    mean()
```


## Batch correction

A very important step for processing SCP data is to correct for batch
effects.  Batch effects are caused by technical variation occurring
during different MS runs. Since only a small number of single-cells
can be acquired at once, batch effects are unavoidable.

The `ComBat()` function from the `sva` package can be used to perform
batch correction as it is performed in the SCoPE2 analysis. We do not
claim that `ComBat` is the best algorithm for batch correcting SCP
data and other batch correcting methods could be used using the same
procedure.

We first extract the assay to process.

```{r get_proteins}
sce <- getWithColData(scp, "proteins_imptd")
```

Next, we need to provide a design matrix and the batch annotation to
`Combat`.  The design matrix allows to protect variables of interest,
in our case `SampleType`.

```{r prepare_batch_correction}
batch <- sce$runCol
model <- model.matrix(~ SampleType, data = colData(sce))
```

We then load and call `ComBat` and overwrite the data matrix. Recall
the data matrix can be accessed using the `assay` function.

```{r compute_batch_correction, results = 'hide', message = FALSE}
library(sva)
assay(sce) <- ComBat(dat = assay(sce),
                     batch = batch,
                     mod = model)
```

Finally, we add the batch corrected assay to the `QFeatures` object
and create the feature links.

```{r add_batch_correction}
scp <- addAssay(scp,
                y = sce,
                name = "proteins_batchC")
scp <- addAssayLinkOneToOne(scp,
                            from = "proteins_imptd",
                            to = "proteins_batchC")
```

For the last time, we plot the overview of the fully processed data
set:

```{r plot5}
plot(scp)
```

# Dimension reduction

Because each assay contains `SingelCellExperiment` objects, we can
easily apply methods developed in the scRNA-Seq field. A useful
package for dimension reduction on single-cell data is the `scater`.

```{r load_scater}
library(scater)
```

This package provides streamline functions to computes various
dimension reduction such as PCA, UMAP, t-SNE, NMF, MDS, ....

## PCA

PCA can be computed using the `runPCA` method. It returns a
`SingleCellExperiment` object for which the dimension reduction
results are stored in the `reducedDim` slot.

```{r run_PCA}
scp[["proteins_batchC"]] <- runPCA(scp[["proteins_batchC"]],
                                   ncomponents = 5,
                                   ntop = Inf,
                                   scale = TRUE,
                                   exprs_values = 1,
                                   name = "PCA")
```

The computed PCA can be displayed using the `plotReducedDim` function. The
`dimred` arguments should give the name of the dimension reduction results to
plot, here we called it `PCA`. The samples are colored by type of sample.

```{r plot_PCA}
plotReducedDim(scp[["proteins_batchC"]],
               dimred = "PCA",
               colour_by = "SampleType",
               point_alpha = 1)
```

This is a minimalistic example with only a few plotted cells, but the
original SCoPE2 dataset contained more than thousand cells.

## UMAP

Similarly to PCA, we can compute a UMAP using the `runUMAP`
method. Note however that the UMAP implementation requires a
initialization, usually provided by PCA. The previous PCA results are
used automatically when supplying `dimred = "PCA"` (`PCA` is the name
of the dimension reduction result that we supplied in the previous
section).

```{r run_UMAP}
scp[["proteins_batchC"]] <- runUMAP(scp[["proteins_batchC"]],
                                    ncomponents = 2,
                                    ntop = Inf,
                                    scale = TRUE,
                                    exprs_values = 1,
                                    n_neighbors = 3,
                                    dimred = "PCA",
                                    name = "UMAP")
```

The computed UMAP can be displayed using the `plotReducedDim`
function. The `dimred` arguments gives the name of the dimension
reduction results to plot, here we called it `UMAP`. The samples are
colored by type of sample.

```{r plot_UMAP}
plotReducedDim(scp[["proteins_batchC"]],
               dimred = "UMAP",
               colour_by = "SampleType",
               point_alpha = 1)
```

The UMAP plot is a very interesting plot for large datasets. A UMAP on
this small example dataset is not useful but is shown for
illustration.

# Monitoring data processing

`QFeatures` keeps the links between the different assays along the
processing of the data. This greatly facilitates the visualization of
the quantitative data for a features at the different processing
levels. For instance, suppose we are interested
in the protein *Plastin-2* (protein ID is `P13796`). A useful QC is to
monitor the data processing at the PSM, peptide and protein
level. This can easily be done thanks to the `QFeatures`
framework. Using the `subsetByFeature`, we can extract the protein of
interest and its related features in the other assays. The data is
formatted to a long format table that can easily be plugged in the
`ggplot2` visualization tool.

```{r monitor_plot, warning = FALSE, fig.width = 10, fig.height = 10, out.width = "100%"}
## Get the features related to Plastin-2 (P13796)
subsetByFeature(scp, "P13796") |>
    ## Format the `QFeatures` to a long format table
    longFormat(colvars = c("runCol", "SampleType", "quantCols")) |>
    data.frame() |>
    ## This is used to preserve ordering of the samples and assays in ggplot2
    mutate(assay = factor(assay, levels = names(scp)),
           Channel = sub("Reporter.intensity.", "TMT-", quantCols),
           Channel = factor(Channel, levels = unique(Channel))) |>
    ## Start plotting
    ggplot(aes(x = Channel, y = value, group = rowname, col = SampleType)) +
    geom_point() +
    ## Plot every assay in a separate facet
    facet_wrap(facets = vars(assay), scales = "free_y", ncol = 3) +
    ## Annotate plot
    xlab("Channels") +
    ylab("Intensity (arbitrary units)") +
    ## Improve plot aspect
    theme(axis.text.x = element_text(angle = 90),
          strip.text = element_text(hjust = 0),
          legend.position = "bottom")
```

This graph helps to keep track of the data processing. We can see how
the different PSMs are progressively aggregated to peptides and then
to proteins as well as how the normalization, imputation or batch
correction impact the distribution of the quantifications.

We have dedicated a separate
[vignette](https://rformassspectrometry.github.io/QFeatures/articles/Visualization.html)
that describes more in details how to visualize and explore data in a
`QFeatures` object.

# Session information {-}

```{r setup2, include = FALSE}
knitr::opts_chunk$set(
    collapse = TRUE,
    comment = "",
    crop = NULL
)
```


```{r sessioninfo, echo=FALSE}
sessionInfo()
```

# License {-}

This vignette is distributed under a [CC BY-SA
license](https://creativecommons.org/licenses/by-sa/2.0/) license.

# Reference {-}