Reading single-cell proteomics data

---

### Get the slides at [https://bit.ly/read_scp_data](https://bit.ly/read_scp_data)

These slides are available under a **creative common
[CC-BY license](http://creativecommons.org/licenses/by/4.0/)**. You are
free to share (copy and redistribute the material in any medium or
format) and adapt (remix, transform, and build upon the material) for
any purpose, even commercially
<img height="20px" alt="CC-BY" src="https://raw.githubusercontent.com/UCLouvain-CBIO/scp-teaching/main/img/cc1.jpg" />.

???
In this presentation, you will learn how to convert data tables into
`QFeatures` objects that can further be used for data processing.

I highly recommend you to watch the previous video about handling
quantitative proteomics features, if that's not already done.

The slides are available at the given link and are shared under CC-BY
license.

---

# How can I convert my single-cell proteomics data to a QFeatures object?

???
So the focus point of this presentation is answering the question:
how can I convert my single-cell proteomics data to a QFeatures object?

---

## How can I convert my single-cell proteomics data to a QFeatures object?

The `readSCP()` function converts quantified mass spectrometry data
tables to `QFeatures` objects.

???
You can do that thanks to the `readSCP()` function. The function
combines an input table and a sample table into a ready-to-process
`QFeatures` object. Let me explain what those two tables correspond to.

---

class:

## Input table

]
.right-column[

Input table = output table from pre-processing software, such as
MaxQuant (*e.g.* `evidence.txt`) or ProteomeDiscoverer (*e.g.*
`PSMs.txt`).

In general, 3 types of columns:

- feature annotations: *e.g.* peptide sequence, ion charge, protein name
- quantification columns: 1 to n (depending on technology)
- acquisition data: *e.g.* file name

]
]
.panel[.panel-name[Example]

<div class="datatables html-widget html-fill-item" id="htmlwidget-4843e7b6b4f9e51ec8ab" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-4843e7b6b4f9e51ec8ab">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8"],["ATNFLAHEK","ATNFLAHEK","SHTILLVQPTK","SHTILLVQPTK","SHTILLVQPTK","SLVIPEK","SLVIPEK","VDREQLVQK"],[9,9,11,11,11,7,7,9],[2,2,2,2,2,2,2,2],[65.78100000000001,63.787,71.884,68.633,71.946,76.20399999999999,78.087,57.329],[0.00052636,0.00058789,4.0315e-24,4.7622e-06,6.8709e-09,0.053705,0.0595,0.004872],["P29692","P29692","P84090","P84090","P84090","P62269","P62269","P61981"],[61251,58648,150350,27347,84035,44895,179560,57895],[501.71,1099.8,3705,405.9,583.09,700.23,5385.4,603.73],[3731.3,2837.7,9361,1525.2,4092.3,2283,13670,2787.9],[1643.3,494.32,0,0,530.13,1109.6,690.53,757.17],[871.84,349.26,1945.4,0,718.13,0,3653.5,862.08],[981.87,1030.5,1188.6,318.74,2204.5,675.79,3624.9,1118.8],[1200.1,0,1574,0,960.51,0,5680.8,640.1],[939.0599999999999,1214.1,2302.1,519.8099999999999,453.77,809.38,5114.5,1446.1],[1457.5,800.58,2176.1,0,1188.4,668.88,4907.3,968.49],[1329.8,807.79,1399.5,507.23,740.99,1467.5,4584,648.5599999999999],[981.83,391.38,1307.5,370.79,0,901.38,4083.4,742.53],[null,null,2192.4,null,null,null,5221.5,null],[null,null,1791.4,null,null,null,4660.6,null],[null,null,1727.5,null,null,null,6841.6,null],[null,null,2157.3,null,null,null,6448.2,null],[null,null,1398,null,null,null,6412.6,null],["190321S_LCA10_X_FP97AG","190222S_LCA9_X_FP94BM","190914S_LCB3_X_16plex_Set_21","190222S_LCA9_X_FP94BM","190321S_LCA10_X_FP97AG","190222S_LCA9_X_FP94BM","190914S_LCB3_X_16plex_Set_21","190321S_LCA10_X_FP97AG"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Sequence<\/th>\n      <th>Length<\/th>\n      <th>Charge<\/th>\n      <th>Retention.time<\/th>\n      <th>PEP<\/th>\n      <th>Leading.razor.protein<\/th>\n      <th>Reporter.intensity.1<\/th>\n      <th>Reporter.intensity.2<\/th>\n      <th>Reporter.intensity.3<\/th>\n      <th>Reporter.intensity.4<\/th>\n      <th>Reporter.intensity.5<\/th>\n      <th>Reporter.intensity.6<\/th>\n      <th>Reporter.intensity.7<\/th>\n      <th>Reporter.intensity.8<\/th>\n      <th>Reporter.intensity.9<\/th>\n      <th>Reporter.intensity.10<\/th>\n      <th>Reporter.intensity.11<\/th>\n      <th>Reporter.intensity.12<\/th>\n      <th>Reporter.intensity.13<\/th>\n      <th>Reporter.intensity.14<\/th>\n      <th>Reporter.intensity.15<\/th>\n      <th>Reporter.intensity.16<\/th>\n      <th>Raw.file<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"scrollX":"200px","paging":false,"info":false,"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Sequence","targets":1},{"name":"Length","targets":2},{"name":"Charge","targets":3},{"name":"Retention.time","targets":4},{"name":"PEP","targets":5},{"name":"Leading.razor.protein","targets":6},{"name":"Reporter.intensity.1","targets":7},{"name":"Reporter.intensity.2","targets":8},{"name":"Reporter.intensity.3","targets":9},{"name":"Reporter.intensity.4","targets":10},{"name":"Reporter.intensity.5","targets":11},{"name":"Reporter.intensity.6","targets":12},{"name":"Reporter.intensity.7","targets":13},{"name":"Reporter.intensity.8","targets":14},{"name":"Reporter.intensity.9","targets":15},{"name":"Reporter.intensity.10","targets":16},{"name":"Reporter.intensity.11","targets":17},{"name":"Reporter.intensity.12","targets":18},{"name":"Reporter.intensity.13","targets":19},{"name":"Reporter.intensity.14","targets":20},{"name":"Reporter.intensity.15","targets":21},{"name":"Reporter.intensity.16","targets":22},{"name":"Raw.file","targets":23}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]
]

???
### Description

The input table is typically a table that is generated by a
pre-processing software, such as MaxQuant or ProteomeDiscoverer.

The input table usually contains 3 types of columns:

- First, some columns hold the feature annotation, think about the peptide
  sequence, the charge of the analyzed ion, the protein name
- You also have columns that hold the quantification data. The number
  of columns may vary depending on the technology and the pre-processing
  software.
- The last type of columns contain data associated to the mass
  spectrometry acquisition. For instance, the name of the file where
  the instrument has stored the data.

Let's have a look at some example data

### Example

The first columns here are annotations of the features, with the
peptide sequence, its length, the ion charge, etc. Then, there is the
quantitative data and you can see all elements are numbers that
correspond to ion intensities. You can identify the quantitative data in this specific
example because the column names start with `Reporter.intensity.`
followed by a number. The last columns is the name of the acquisition
and relates to the mass spectrometry run.

---

class:

## Sample annotation

]
.right-column[

Sample table = table generated by the researcher.

1 line = 1 sample

Two columns are **required**:

- Names of the quantification columns from the input table
- Acquisition name, same as in the input table

Other columns can contain additional sample annotations, such as:

- Experiment metadata (date, researcher's name, instruments, ...)
- Sample preparation (cell culture batch, LC batch, TMT label, ...)
- Sample metadata (species, treatment, disease, sex, age, ...)
- Sample type (single-cells, carrier, blanks, ...)
- Other data (FACS data, microscopy data, phenotypic data, ...)
- ...

]
]
.panel[.panel-name[Example]

<div class="datatables html-widget html-fill-item" id="htmlwidget-d13738759baa9ea2b858" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-d13738759baa9ea2b858">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6","7","8"],["190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM"],["Reporter.intensity.1","Reporter.intensity.2","Reporter.intensity.3","Reporter.intensity.4","Reporter.intensity.5","Reporter.intensity.6","Reporter.intensity.7","Reporter.intensity.8"],["Carrier","Reference","Unused","Monocyte","Blank","Monocyte","Macrophage","Macrophage"],["LCA9","LCA9","LCA9","LCA9","LCA9","LCA9","LCA9","LCA9"],["s8","s8","s8","s8","s8","s8","s8","s8"],["N","N","N","N","N","N","N","N"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Raw.file<\/th>\n      <th>Channel<\/th>\n      <th>SampleType<\/th>\n      <th>lcbatch<\/th>\n      <th>sortday<\/th>\n      <th>digest<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"scrollX":"200px","paging":false,"info":false,"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Raw.file","targets":1},{"name":"Channel","targets":2},{"name":"SampleType","targets":3},{"name":"lcbatch","targets":4},{"name":"sortday","targets":5},{"name":"digest","targets":6}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]
]

???
### Description

Next to the input table, the `readSCP` function requires a sample table.
This is generated by the researcher. Each line contains information
about a single sample.

2 columns are required to work with `readSCP`. The first column tells
the software what are the names of the columns containing the
quantification data in the input table. The other column contains the
acquisition names, just like we saw for the input table.

Beside those 2 required columns, you can include any sample data you may
have, be it experimental metadata, sample preparation information,
sample types or other data collected during sample preparation. These
data are valuable when it comes to data modelling!

### Example

In this example, the 2 first columns are the required columns. The
first column is very similar to the column in the input data and may
notice that the second column holds the names that I pointed out as
quantitative columns in the input table that were starting with
`Reporter.intensity.`. The remaining columns are examples of
additional information that could be available.

---

# What happens under the hood?

???
You may ask yourself: what is `readSCP` exactly doing?

---

class:

## What happens under the hood?

]
.right-column[

The feature annotations are separated from the quantitative data.

The two table pieces are converted to a
[`SingleCellExperiment`](https://www.bioconductor.org/packages/release/bioc/vignettes/SingleCellExperiment/inst/doc/intro.html)
object [1], a specialized **Bioconductor** data container that creates an interface to
existing functions to analyse single-cell data.

<br><br><br>
<p style="color:grey;font-size:0.75em;">
[1] Amezquita, Robert A., Aaron T. L. Lun, Etienne Becht, Vince J. Carey, Lindsay N. Carpp, Ludwig Geistlinger, Federico Martini, et al. 2019. “Orchestrating Single-Cell Analysis with Bioconductor.” Nature Methods, December, 1–9.
</p>

]
]
.panel[.panel-name[Step2]

]
.right-column[

The data is then split based on the acquisition name.

Each quantitative column now corresponds to a **single** and **unique**
sample.

]
]
.panel[.panel-name[Step3]
.pull-left[

]
.pull-right[

The sample table is matched to the split feature data. This is
performed based on the two **required** columns:

- Acquisition name to match each data piece
- Quantification column names to match columns in each data piece

**Unique sample IDs** are created

]
]
.panel[.panel-name[Step4]
.pull-left[

]
.pull-right[

All the data pieces are wrapped into a `QFeatures` object.

Overall, the `QFeatures` format enables seamless **data management and access**,
important for **downstream** data processing and visualisation.
]
]
]

???
Well, `readSCP` prepares the data in 4 main steps.

### Step1

First, it takes the input table and separates the feature annotation
from the quantitative data. The two tables can then be converted to a
SingleCellExperiment object that is a specialized Bioconductor data
container that creates an interface to existing functions to analyse
single-cell data.

### Step2

Next, the data is further split, but this time, the split is along the
rows based on the mass run. Because the
acquisitions are now separated, each quantitative column now corresponds
to a single and unique sample.

### Step3

In step3, the sample table comes in. The sample annotations are linked
to the quantification data and this is performed based on the two required columns I just described.
First, the acquisition names are matched between the input table and
the sample table. Next, the quantitative column names from the sample
tables are used to link each sample to its corresponding quantification
column. The combination of quantification column name and acquisition
name creates unique sample identifiers that are stored in the sample
table.

### Step4

The final step is to wrap all those data pieces into a `QFeatures`
object. Overall, the QFeatures format enables seamless data management
and access that are important for downstream data processing and
visualisation.

---

# readSCP() in practice

???
Let me now show you how to use `readSCP` in practice

---

class:

## `readSCP()` in practice

`sampleTable`

<div class="datatables html-widget html-fill-item" id="htmlwidget-8983f24dc9e24333a000" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-8983f24dc9e24333a000">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM","190222S_LCA9_X_FP94BM"],["Reporter.intensity.1","Reporter.intensity.2","Reporter.intensity.3","Reporter.intensity.4","Reporter.intensity.5","Reporter.intensity.6"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Raw.file<\/th>\n      <th>Channel<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"scrollX":"200px","paging":false,"info":false,"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Raw.file","targets":1},{"name":"Channel","targets":2}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]
.pull-right[

`inputTable`

<div class="datatables html-widget html-fill-item" id="htmlwidget-e48d0b10bae63d708716" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-e48d0b10bae63d708716">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["190321S_LCA10_X_FP97AG","190222S_LCA9_X_FP94BM","190914S_LCB3_X_16plex_Set_21","190222S_LCA9_X_FP94BM","190321S_LCA10_X_FP97AG","190222S_LCA9_X_FP94BM"],[61251,58648,150350,27347,84035,44895],[501.71,1099.8,3705,405.9,583.09,700.23],[3731.3,2837.7,9361,1525.2,4092.3,2283],[1643.3,494.32,0,0,530.13,1109.6],[871.84,349.26,1945.4,0,718.13,0],[981.87,1030.5,1188.6,318.74,2204.5,675.79],[1200.1,0,1574,0,960.51,0],[939.0599999999999,1214.1,2302.1,519.8099999999999,453.77,809.38],[1457.5,800.58,2176.1,0,1188.4,668.88],[1329.8,807.79,1399.5,507.23,740.99,1467.5],[981.83,391.38,1307.5,370.79,0,901.38],[null,null,2192.4,null,null,null],[null,null,1791.4,null,null,null],[null,null,1727.5,null,null,null],[null,null,2157.3,null,null,null],[null,null,1398,null,null,null]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>Raw.file<\/th>\n      <th>Reporter.intensity.1<\/th>\n      <th>Reporter.intensity.2<\/th>\n      <th>Reporter.intensity.3<\/th>\n      <th>Reporter.intensity.4<\/th>\n      <th>Reporter.intensity.5<\/th>\n      <th>Reporter.intensity.6<\/th>\n      <th>Reporter.intensity.7<\/th>\n      <th>Reporter.intensity.8<\/th>\n      <th>Reporter.intensity.9<\/th>\n      <th>Reporter.intensity.10<\/th>\n      <th>Reporter.intensity.11<\/th>\n      <th>Reporter.intensity.12<\/th>\n      <th>Reporter.intensity.13<\/th>\n      <th>Reporter.intensity.14<\/th>\n      <th>Reporter.intensity.15<\/th>\n      <th>Reporter.intensity.16<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"scrollX":"200px","paging":false,"info":false,"columnDefs":[{"className":"dt-right","targets":[2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"Raw.file","targets":1},{"name":"Reporter.intensity.1","targets":2},{"name":"Reporter.intensity.2","targets":3},{"name":"Reporter.intensity.3","targets":4},{"name":"Reporter.intensity.4","targets":5},{"name":"Reporter.intensity.5","targets":6},{"name":"Reporter.intensity.6","targets":7},{"name":"Reporter.intensity.7","targets":8},{"name":"Reporter.intensity.8","targets":9},{"name":"Reporter.intensity.9","targets":10},{"name":"Reporter.intensity.10","targets":11},{"name":"Reporter.intensity.11","targets":12},{"name":"Reporter.intensity.12","targets":13},{"name":"Reporter.intensity.13","targets":14},{"name":"Reporter.intensity.14","targets":15},{"name":"Reporter.intensity.15","targets":16},{"name":"Reporter.intensity.16","targets":17}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]
]
.panel[.panel-name[Code]

```r
readSCP(inputTable,
        sampleTable,
        batchCol = "Raw.file",
        channelCol = "Channel")
```

Overview of the resulting `QFeatures` object:

```
## An instance of class QFeatures containing 4 assays:
##  [1] 190222S_LCA9_X_FP94BM: SingleCellExperiment with 395 rows and 16 columns 
##  [2] 190321S_LCA10_X_FP97_blank_01: SingleCellExperiment with 109 rows and 16 columns 
##  [3] 190321S_LCA10_X_FP97AG: SingleCellExperiment with 487 rows and 16 columns 
##  [4] 190914S_LCB3_X_16plex_Set_21: SingleCellExperiment with 370 rows and 16 columns
```

]
]

???
### Data

Consider the two example tables here that are very similar to the ones I
showed previously. Note that both table contain the `Raw.file` column
necessary for mathcing the acquisition runs. See also that the `Channel`
column in the `sampleTable` contains the column names corresponding to
the quantification columns in the `inputTable`.

### Code

You can convert those two tables to a `QFeatures` object using
`readSCP()` by running the command shown here. We call `readSCP()`, we
provide the 2 tables, but we also need to tell how the matching is
performed. We here tell the function which column in both tables
should be used to match the acquisition batch, in this case remember
it was `Raw.file`. Then, we need to tell the function which column in
the sample table contains the quantification columns, it's the column
called `Channel` in this example.

When running the command, we create a new object and you can find an
overview here. You can see that we indeed have a `QFeatures` object with 4 assays
because there are 4 acquisitions. Each assay is a SingleCellExperiment
object with variable number of rows, here corresponding to different
quantified PSMs and 16 columns because a multiplexing with 16 labels
was used to acquire the data.

---

# What if I acquired a single acquisition or already processed my data?

???
As a mention, you can also import data from a single acquisition or
import already processed single-cell proteomics data.

---

## What if I acquired a single acquisition or already processed my data?

`readQFeatures()`: import pre-processed data for a single acquisition
as a `QFeatures` object.

```r
readQFeatures(inputTable, ecol = paste0("Reporter.intensity.", 1:16),
              colData = sampleTable)
```

`readSingleCellExperiment()`: import data as a `SingleCellExperiment`
object.

```r
readSingleCellExperiment(inputTable, ecol = paste0("Reporter.intensity.", 1:16),
                         colData = sampleTable)
```

Both functions require:
 - input table
 - quantification column names

Optional: sample annotation table.

???
Suppose you acquired data from a single acquisition, you may not want
to bother with splitting your data using `readSCP()` but rather use
the simpler function `readQFeatures()`.

Also, if you have already processed your data or if you downloaded data
ready to analyse, you may simply want to load it as a
`SingleCellExperiment` object for downstream analyses (we will come to
that in the next slide deck). You can do so using the `readSingleCellExperiment()`.

Both functions require an input table, similar to `readSCP` and you need
to point out which columns are the quantification columns. Optionally, you
can provide sample annotations.

I hope this presentation helped you understand how to get your
single-cell proteomics data into R for data processing and analysis
using our software.

---

# Exercise

???
I suggest to test your understanding with a small exercise.

---

#### Given the input and sample tables, what command creates a single-cell proteomics QFeatures object?

`inputTable`

<div class="datatables html-widget html-fill-item" id="htmlwidget-efa3ce13d441eff055fe" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-efa3ce13d441eff055fe">{"x":{"filter":"none","vertical":false,"data":[["1","2","3"],["A","A","B"],["AYGGEGP","QSRTSGRG","QSRTSGRG"],[4829,6166,6862],[2346,6599,2834]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>MSrun<\/th>\n      <th>Peptide<\/th>\n      <th>lab1<\/th>\n      <th>lab2<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"paging":false,"searching":false,"info":false,"columnDefs":[{"className":"dt-right","targets":[3,4]},{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"MSrun","targets":1},{"name":"Peptide","targets":2},{"name":"lab1","targets":3},{"name":"lab2","targets":4}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]
.pull-right[

`sampleTable`

<div class="datatables html-widget html-fill-item" id="htmlwidget-026cf3a7bf8f40453c6a" style="width:100%;height:auto;"></div>
<script type="application/json" data-for="htmlwidget-026cf3a7bf8f40453c6a">{"x":{"filter":"none","vertical":false,"data":[["1","2","3","4","5","6"],["A","A","A","A","B","B"],["lab1","lab2","lab1","lab2","lab1","lab2"],["HeLa","HEK293","HeLa","HEK293","HeLa","HEK293"]],"container":"<table class=\"display\">\n  <thead>\n    <tr>\n      <th> <\/th>\n      <th>MSrun<\/th>\n      <th>Label<\/th>\n      <th>CellType<\/th>\n    <\/tr>\n  <\/thead>\n<\/table>","options":{"paging":false,"searching":false,"info":false,"columnDefs":[{"orderable":false,"targets":0},{"name":" ","targets":0},{"name":"MSrun","targets":1},{"name":"Label","targets":2},{"name":"CellType","targets":3}],"order":[],"autoWidth":false,"orderClasses":false}},"evals":[],"jsHooks":[]}</script>

]

```r
1. readSCP(inputTable, sampleTable, batchCol = "Label",  channelCol = c("lab1", "lab2"))
2. readSCP(inputTable, sampleTable, batchCol = "MSrun",  channelCol = c("lab1", "lab2"))
3. readSCP(inputTable, sampleTable, batchCol = "MSrun",  channelCol = "Label")
4. readSCP(inputTable, sampleTable, batchCol = "CellType",  channelCol = "Label")
```

???
Given the input and sample tables, What command creates a single-cell
proteomics QFeatures object?

Command 1, 2, 3 or 4? You can pause the video to think about it... The
solution is command 3. See that we can match the acquisition runs using
the `MSrun` column in both tables and the `Label` column in the sample
table contains the quantification column names of the input table.

---

### Further information

Learn more about loading single-cell proteomics data as a `QFeatures`
object in our dedicated vignette at
https://uclouvain-cbio.github.io/scp/articles/read_scp.html.

### Funding

Fonds de la Recherche Scientifique (FNRS), Belgium

???
I hope you found the correct solution. If you're looking for more
detailed information, you can have a look at our vignette dedicated to
loading single-cell proteomics data. Thank you very much for watching
and see you in another video.