concept · intermediate · partly written
mRNA-protein correlation
How well mRNA level predicts protein level depends on the axis, the noise treatment and the statistic.
Also called: transcript protein correlation
The short answer
There is no single mRNA-protein correlation. It depends on whether you compare genes within a sample or one gene across samples, whether noise is corrected, and which correlation statistic is used. 12
Reported, not yet replicated
Reported correlations, with their conditions more
Human, across genes, one line: Pearson 0.54 (R2 0.29) over 512 high-confidence genes of 1,025 matched, Daoy medulloblastoma, 0.57 after reliability correction; Spearman 0.6, HeLa, 10,255 proteins. Human, targeted MS, 55 chosen genes, 9 cell lines and 11 tissues: median per-sample Pearson 0.67 (HEK293 0.39, MCF7 0.79), 0.93 after a per-gene factor estimated from the other samples. Human tissues: a tissue's own mRNA explains 20% (ovary) to 39% (liver) of protein variance across genes. Mouse tissues: Spearman about 0.5 to 0.6 for matched tissue, 0.1 to 0.3 for unrelated tissue. One gene across samples: mean r 0.29 (12 human tissues); per-gene medians 0.42 (949 cell lines), 0.31 (134 lines), 0.40 (tumours); mean 0.48 for 375 CCLE lines (unverified in the report). Yeast, 24 studies: raw r 0.618, after noise model 0.926. Early: Anderson and Seilhamer 1997 reported r = 0.48 for 19 human liver genes. 345671891011
Reported, not yet replicated
What accounts for the gap more
Four rates set steady state: mRNA made, mRNA removed, protein made, protein removed; RNA-seq sees only the first two folded together. Ribosome profiling in yeast found translation efficiency varying about 100-fold between genes; in 29 human tissues the protein-to-mRNA ratio spans over two orders of magnitude and a sequence-based model predicts it to a median 3.2-fold. Buffering also contributes: eQTL effects were significantly reduced at the protein level. 12613
Reported, not yet replicated
How much is measurement noise more
Yeast: most of the gap (24 studies; averaging replicates raised apparent variance explained from 33% to 72%; r 0.926 after correction). One human cell line: little (0.54 to 0.57). Human tissues: reliabilities of 0.77 (mRNA) and 0.21 (protein) imply about half of across-tissue protein variance is mRNA after correction; the authors flag large uncertainty and bias from shared replicate error. Cancer panels: replicate reproducibility explains about three times more variance in per-gene correlation than complex membership. 10312
Reported, not yet replicated
Where studies disagree more
Mouse NIH3T3: one analysis says translation dominates; a re-scaling of the same data (61 housekeeping proteins) says mRNA explains at least 56% to about 81%. The authors of the second call their numbers tentative. A third study in dendritic cells puts mRNA at 59 to 68% at steady state and 87 to 92% of fold changes after stimulation. A 2013 correction to the first paper exists and its text was not opened. 141516
Contested
The two ends of the debate more
A 2016 review (Liu, Beyer, Aebersold) concludes transcript levels alone often cannot predict protein levels; a 2012 review (Vogel and Marcotte) concludes post-transcriptional regulation substantially shapes protein abundance. Both were read at abstract or record level only. The yeast noise-corrected result is not in conflict because it is a different axis; that reconciliation is our inference. 171810
Our inference
Plain answer
Across genes in one human sample the correlation is moderate (about 0.4 to 0.8 in the studies read); for one gene across samples it is weaker (mean 0.29 across 12 tissues); in yeast, after a noise model, it reaches 0.93. The gap is part biology and part measurement, and the split is unsettled. 3110
Reported, not yet replicated
Why it matters
Gene-expression data (RNA-seq) are often used as a stand-in for protein. How good the stand-in is depends on the question. 1
Our inference
Equation
protein level = mRNA level x translation rate / protein removal rate (steady state)
Real numbers
Yeast, noise-corrected
r 0.926 (86% of variance), 24 studies, up to 5,854 genes. 10
Reported, not yet replicated
Still open here
- Is there a noise-corrected human figure for one cell type with matched RNA and protein replicates?
- Is the Franks 0.70 after per-gene scaling (shown only in a figure) correct?
Linked from
- Mass-spectrometry proteomics
- RNA-seq
- How omics layers are joined: by identifier, sample, cell, position, latent space and graph
- Models of models
- Proteomics standards and archives: mzML, mzTab, ProForma, SDRF and ProteomeXchange
- RNA and protein half-lives in human cells: what has been measured
- Translation initiation: scanning and start-codon choice
- Matthias Mann
- Ruedi Aebersold
- How much of the human mRNA-protein gap is measurement rather than biology?
Sources
- Franks A, Airoldi E, Slavov N. Post-transcriptional regulation across human tissues. PLoS Comput Biol (2017). paper; opened 2026-11-17 all records citing this
- Upadhya SR, Ryan CJ. Experimental reproducibility limits the correlation between mRNA and protein abundances in tumor proteomic profiles. Cell Rep Methods (2022). paper; opened 2026-11-17 all records citing this
- Vogel C, Abreu Rde S, Ko D, et al.. Sequence signatures and mRNA concentration can explain two-thirds of protein abundance variation in a human cell line. Mol Syst Biol (2010). paper; opened 2026-11-17 all records citing this
- Nagaraj N, Wisniewski JR, Geiger T, et al.. Deep proteome and transcriptome mapping of a human cancer cell line. Mol Syst Biol (2011). paper; opened 2026-11-17 all records citing this
- Edfors F, Danielsson F, Hallström BM, et al.. Gene-specific correlation of RNA and protein levels in human cells and tissues. Mol Syst Biol (2016). paper; opened 2026-11-17 all records citing this
- Eraslan B, Wang D, Gusic M, et al.. Quantification and discovery of sequence determinants of protein-per-mRNA amount in 29 human tissues. Mol Syst Biol (2019). paper; opened 2026-11-17 all records citing this
- Geiger T, Velic A, Macek B, et al.. Initial quantitative proteomic map of 28 mouse tissues using the SILAC mouse. Mol Cell Proteomics (2013). paper; opened 2026-11-17 all records citing this
- Gonçalves E, Poulos RC, Cai Z, et al.. Pan-cancer proteomic map of 949 human cell lines. Cancer Cell (2022). paper; opened 2026-11-17 all records citing this
- Jarnuczak AF, Najgebauer H, Barzine M, et al.. An integrated landscape of protein expression in human cancer. Sci Data (2021). paper; opened 2026-11-17 all records citing this
- Csárdi G, Franks A, Choi DS, et al.. Accounting for experimental noise reveals that mRNA levels, amplified by post-transcriptional processes, largely determine steady-state protein levels in yeast. PLoS Genetics (2015). paper; opened 2026-11-17 all records citing this
- Anderson L, Seilhamer J. A comparison of selected mRNA and protein abundances in human liver. Electrophoresis (1997). paper; opened 2026-11-17 all records citing this
- Ingolia NT, Ghaemmaghami S, Newman JR, et al.. Genome-wide analysis in vivo of translation with nucleotide resolution using ribosome profiling. Science (2009). paper; opened 2026-11-17 all records citing this
- Battle A, Khan Z, Wang SH, et al.. Genomic variation. Impact of regulatory variation from RNA to protein. Science (2015). paper; opened 2026-11-17 all records citing this
- Schwanhäusser B, Busse D, Li N, et al.. Global quantification of mammalian gene expression control. Nature (2011). paper; opened 2026-11-17 all records citing this
- Li JJ, Bickel PJ, Biggin MD. System wide analyses have underestimated protein abundances and the importance of transcription in mammals. PeerJ (2014). paper; opened 2026-11-17 all records citing this
- Jovanovic M, Rooney MS, Mertins P, et al.. Dynamic profiling of the protein life cycle in response to pathogens. Science (2015). paper; opened 2026-11-17 all records citing this
- Liu Y, Beyer A, Aebersold R. On the Dependency of Cellular Protein Levels on mRNA Abundance. Cell (2016). review; opened 2026-11-17 all records citing this
- Vogel C, Marcotte EM. Insights into the regulation of protein abundance from proteomic and transcriptomic analyses. Nat Rev Genet (2012). review; opened 2026-11-17 all records citing this
Where it sits on the map
History of this record (1)
This record has been changed 1 time, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.
- Status
- partly written
- Last updated
- 2026-11-17
- Record
- content/concepts/mrna-protein-correlation.json