Fig 1.
Transcriptome and proteome integration in human CD8+ T cells.
(A, B) mRNA abundance in transcript per kilobase per million (TPM) and protein abundance data in protein copy number (CN) of CD8+ T cells that were activated for 2 days with αCD3/αCD28, cultured for 4 days, and were then re-activated for 4h with PMA-Ionomycin (TActivated; n = 12; from [23]). (C) Estimated protein mass per cells using the proteomic ruler methodology [25] of blood-derived, naïve (TN; CD45RA+ CD62Lhigh; n = 4), central memory (TCM; CD45RO+ CD62Lhigh CX3CR1-; n = 4) and effector-memory (TEM; CD45RO+, CD62Llow, CX3CR1high; n = 4) CD8+ T cell subsets and for CD8+ T cells from (A). The p-value resulting from two-tailed t-test is indicated in (C). (D) Graph representing mRNA and protein correlation (such as Pearson’s correlation) and slope of a linear regression. Of note, the correlation coefficient represents a numerical integration of the dispersion of the data points and is not a measure of the angle of the regression line. (E) Integrated mRNA and protein abundance in TActivated. The Pearson’s correlation coefficient, the linear regression (grey line), and its mRNA-to-protein slope (also known as Beta-coefficient) are indicated in the graph. (F) Table indicating the Pearson’s correlation coefficient and the mRNA-to-protein slope for T cell populations in (C).
Fig 2.
Gene class-specific mRNA and protein abundance.
(A) Integrated mRNA (TPM) and protein abundance (CN) of secreted proteins, CD molecules, transcription factors, RNA-binding proteins (RBPs), mammalian (light green) and mitochondrial (dark green) ribosomal proteins, and TCA (citric acid) cycle proteins of activated CD8+ T cells (TActivated). The number of genes per class, the Pearson’s correlation coefficient, and regression line (linear model) were calculated for each class and are indicated in the panels. The correlation was performed independently for mammalian (light green) and mitochondrial (dark green) ribosomal proteins. (B, C) Pearson’s correlation coefficient (B) and mRNA-to-protein slope (also known as Beta-coefficient) (C) of the integrated mRNA and protein abundance of gene classes indicated in (A) during T cell differentiation and upon T cell activation. “*” indicates correlation and regression with p-value <0.05.
Fig 3.
Distinct gene characteristics are associated with gene functions.
(A-D) Relation between gene classes (Fig 2) and sequence homology of human genes with Zebrafish (D. rerio) (A), the length of the 5’UTR (B) and 3’UTR (C), and the GC content (D). Red bars indicate the median values per gene class, which are also indicated numerically at each respective graph. nt: nucleotides.
Fig 4.
Altered mRNA and protein abundance is associated with gene characteristics.
(A, B) The effect of sequence conservation on mRNA (A, left panel) and protein abundance (A, right panel) and the relation thereof (B) in PMA-Ionomycin activated CD8+ T cells. Genes products were stratified according to S3 Fig, in: (A, B) high (>76.9%), intermediate (between 50–76.9%) and low (<50%) orthologous sequence conservation between Human and Zebra-fish (Danio rerio); (C, D) short (<~50nt), intermediate (~50nt- ~200nt), and long 5’UTRs (>~200nt); (E, F) short (<~100nt), intermediate (~100- ~1,000nt), and long (>~1,000nt) 3’UTRs; (G, H) high (>50.9%), intermediate (39.7%<GC<50.9%) and low (<39.7%) GC content per gene. The slope of the regression line of each group is color-coded in B, D, F, H. Differences between groups were assessed with a two-tailed t-test followed by adjustment of p-value using the Benjamini-Hochberg procedure. Adjusted p-value are indicated. Only gene products that are detected both at mRNA and protein levels are shown. The mean per group (red line) is indicated in A,C,E,G. TPM: Transcripts per kilo-base per million; CN: Protein copy number. “*” indicates correlation and regression with p-value <0.05.
Fig 5.
ARE clusters differentially modulate mRNA and protein levels in resting and activated CD8+ T cells.
(A-B) mRNA (TPM) and protein (CN) abundance in (A) CD45RO+ CD8+ memory T cell subsets (TMEM; n = 16; from [22]) and (B) TActivated CD8+ T cells of all genes, grouped according to the number of AU-rich element in cluster within the transcript’s 3’UTR clusters according to ARED-plus annotation (A; see methods). Differences between groups were assessed with a two-tailed t-test followed by adjustment of p-value using the Benjamini-Hochberg procedure. Adjusted p-value are indicated. (C) and integrated mRNA and protein abundance. The slope of the regression line of each group is color-coded in (A, B), and only gene products that are detected both at mRNA and protein levels are shown. TPM: Transcripts per kilo-base per million; CN: Protein copy number. n.s.: not significant. “*” indicates correlation and regression with p-value <0.05.
Fig 6.
ARE cluster content differentially affect gene classes.
(A-F) Integrated mRNA (TPM) and protein (CN) abundance of genes co-detected at mRNA and protein levels, for (A, B) CD molecules, (C, D) transcription factors, or (E, F) secreted proteins. Genes were stratified according to the number of AU-rich element in cluster within the transcript’s 3’UTR, for (A, C, E) TMEM or (B, D, F) TActivated CD8+ T cells. The linear regression and its slope are color-coded for each group. Difference between groups were assessed with a two-tailed t-test followed by adjustment of p-value using the Benjamini-Hochberg procedure. Adjusted p-value are indicated. TPM: Transcripts per kilo-base per million; CN: Protein copy number. “*” indicates correlation and regression with p-value <0.05.