dou2019_boosting.RdSingle-cell proteomics using nanoPOTS combined with TMT isobaric labeling. It contains quantitative information at PSM and protein level. The cell types are either "Raw" (macrophage cells), "C10" (epihelial cells), or "SVEC" (endothelial cells). Each cell is replicated 2 or 3 times. Each cell type was run using 3 levels of boosting: 0 ng (no boosting), 5 ng or 50 ng. When boosting was applied, 1 reference well and 1 boosting well were added, otherwise 1 empty well was added. Each boosting setting (0ng, 5ng, 50ng) was run in duplicate.
dou2019_boostingA QFeatures object with 7 assays, each assay being a SingleCellExperiment object:
Boosting_X_run_Y: PSM data with 10 columns corresponding to
the TMT-10plex channels. The X indicates the boosting amount
(0ng, 5ng or 50ng) and Y indicates the run number (1 or 2).
peptides: peptide data containing quantitative data for 13,462
peptides in 60 samples (run 1 and run 2 combined).
proteins: protein data containing quantitative data for 1436
proteins and 60 samples (all runs combined).
Sample annotation is stored in colData(dou2019_boosting()).
The PSM data can be downloaded from the massIVE repository MSV000084110. FTP link: ftp://massive.ucsd.edu/MSV000084110/
The protein data can be downloaded from the ACS Publications website (Supporting information section).
The data were acquired using the following setup. More information
can be found in the source article (see References).
Cell isolation: single-cells from the three murine cell lines were isolated using FACS (BD Influx II cell sorter ). Boosting sample were prepared (presumably in bulk) from 1:1:1 mix of the three cell lines.
Sample preparation performed using the nanoPOTs device. Protein extraction (DMM + TCEAP) + alkylation (IAA) + Lys-C digestion + trypsin digestion + TMT-10plex labeling and pooling.
Separation: nanoLC (Dionex UltiMate with an in-house packed 50cm x 30um LC columns; 50nL/min)
Ionization: ESI (2,000V)
Mass spectrometry: Thermo Fisher Orbitrap Fusion Lumos Tribrid (MS1 accumulation time = 50ms; MS1 resolution = 120,000; MS1 AGC = 1E6; MS2 accumulation time = 246ms; MS2 resolution = 60,000; MS2 AGC = 1E5)
Data analysis: MS-GF+ + MASIC (v3.0.7111) + RomicsProcessor (custom R package)
The PSM data were collected from the MassIVE repository
MSV000084110 (see Source section). The downloaded files are:
Boosting_*ng_run_*_msgfplus.mzid: the MS-GF+ identification
result files.
Boosting_*ng_run_*_ReporterIons.txt: the MASIC quantification
result files.
For each batch, the quantification and identification data were
combined based on the scan number (common to both data sets). The
combined datasets for the different runs were then concatenated
feature-wise. To avoid data duplication due to ambiguous matching
of spectra to peptides or ambiguous mapping of peptides to proteins,
we combined ambiguous peptides to peptides groups and proteins to
protein groups. Feature annotations that are not common within a
peptide or protein group are are separated by a ;. The sample
annotation table was manually created based on the available
information provided in the article. The data were then converted
to a QFeatures object using the scp::readSCP() function.
We generated the peptide data. First, we removed PSM matched to contaminants or decoy peptides and ensured a 1% FDR. We aggregated the PSM to peptides based on the peptide (or peptide group) sequence(s) using the median PSM instenity. The peptide data for the different runs were then joined in a single assay (see QFeatures::joinAssays), again based on the peptide sequence(s). We then removed the peptide groups. Links between the peptide and the PSM data were created using QFeatures::addAssayLink. Note that links between PSM and peptide groups are not stored.
The protein data were downloaded from Supporting information
section from the publisher's website (see Sources). The data is
supplied as an Excel file ac9b03349_si_004.xlsx. The file
contains 7 sheets from which we took the 2nd, 4th and 6th sheets
(named 01 - No Boost raw data, 03 - 5ng boost raw data,
05 - 50ng boost raw data, respectively). The sheets contain the
combined protein data for the duplicate runs given the boosting
amount. We joined the data for all boosting ration based on the
protein name and converted the data to a SingleCellExperiment
object. We then added the object as a new assay in the QFeatures
dataset (containing the PSM data). Links between the proteins and
the corresponding PSM were created. Note that links to protein
groups are not stored.
Dou, Maowei, Geremy Clair, Chia-Feng Tsai, Kerui Xu, William B. Chrisler, Ryan L. Sontag, Rui Zhao, et al. 2019. “High-Throughput Single Cell Proteomics Enabled by Multiplex Isobaric Labeling in a Nanodroplet Sample Preparation Platform.” Analytical Chemistry, September (link to article).
# \donttest{
dou2019_boosting()
#> see ?scpdata and browseVignettes('scpdata') for documentation
#> loading from cache
#> An instance of class QFeatures containing 8 assays:
#> [1] Boosting_0ng_run_1: SingleCellExperiment with 12258 rows and 10 columns
#> [2] Boosting_0ng_run_2: SingleCellExperiment with 12611 rows and 10 columns
#> [3] Boosting_50ng_run_1: SingleCellExperiment with 27746 rows and 10 columns
#> ...
#> [6] Boosting_5ng_run_2: SingleCellExperiment with 20435 rows and 10 columns
#> [7] peptides: SingleCellExperiment with 13462 rows and 60 columns
#> [8] proteins: SingleCellExperiment with 1436 rows and 60 columns
# }