Fig. 7: IceR boosts the completeness and quality of single-cell proteomics data.
From: IceR improves proteome coverage and data completeness in global and single-cell proteomics
a Numbers of proteins (upper panel) and peptides (lower panel) and their average (dashed line) quantified by MaxQuant per single hair (FM1-43high, green) and progenitor (FM1-43low, grey) cell sample. b As in a, but after data reprocessing by IceR. c Fraction of missing values on protein-level and peptide-level in MaxQuant (grey) and IceR (orange) results. d Dimensional reduction of protein abundance estimations in MaxQuant data by t-distributed stochastic neighbour embedding (tSNE). Progenitor cells (FM1-43low) are coloured in grey. Hair cells (FM1-43high) are coloured in green. Silhouette score (S) is indicated. e As in d but after data reprocessing by IceR. f Volcano plot showing detected significantly (orange, adj. p value < 0.05, absolute fold-change > 2) differently abundant proteins between single hair cells and single progenitor cells in IceR data. Significance cut-off is indicated by a dashed line. g Chronological ordering of single-cells as a function of CellTrails’ inferred pseudotime from IceR data. Single-cells are coloured according to their FM1-43 uptake. h Scaled expression dynamics over pseudotime for all analysed proteins in IceR data based on generalized additive models (GAM). Low and high temporal protein expression is indicated by blue and red colour tones, respectively. i Absolute expression dynamics of log2 expression levels as a function of pseudotime for various proteins. Single-cells are coloured according to their FM1-43 uptake. j Ion density in IceR-selected DICE windows per single-cell sample (upper panels) for the peptides VLTLDLYK and AGATTIEAVK of OTOF (hair cell marker protein) and TPM3 (progenitor cell marker protein), respectively. Corresponding peptide abundances (lower panels) are ordered by CellTrails‘ inferred pseudotime.
