2D-DIGE vs LC-MS/MS: Strengths, Limitations and When to Use Each
Two-dimensional difference gel electrophoresis separates intact proteins and shows you a map; shotgun LC-MS/MS digests everything to peptides and identifies thousands of proteins in a run. They answer different questions, and the honest comparison is not which is better but which is right for the question you have. This page sets out what each does well, what each misses, and why many labs use both.
What each method does
2D-DIGE separates intact proteins by isoelectric point and molecular weight, with samples pre-labeled by fluorescent dyes so that up to three can share a gel. The output is an image: a map of spots, each spot a protein or proteoform, with intensity proportional to abundance. Differences are found by comparing spot intensities across gels, and the spots that differ are then cut out and identified by mass spectrometry. It is a top-down separation with a bottom-up identification step.
Shotgun LC-MS/MS, also called bottom-up proteomics, digests the whole sample to peptides, separates the peptides by liquid chromatography, and identifies and quantifies them by tandem mass spectrometry. Proteins are inferred from their peptides. The output is a list: thousands of proteins with abundance values, no image, and, because the proteins were digested before analysis, no direct view of the intact protein.
Depth: how many proteins
On depth, LC-MS/MS wins and it is not close. A modern shotgun run identifies thousands of proteins from a complex sample. A modern 2D gel resolves around 2,500 to 3,000 spots, which, at an average of about three spots per protein, corresponds to roughly 800 gene products. Larger formats and prefractionation push that up, and gels have been reported resolving more than 10,000 spots corresponding to more than 1,000 proteins, but the gap remains.
What the number hides is which proteins. The 800 or so proteins a 2D gel resolves are the abundant ones, and shotgun data show that they account for about 90% of the protein mass of a cell. A gel therefore sees most of the proteome by weight and a minority of it by count. Whether that matters depends entirely on whether the proteins you care about are abundant.
Proteoforms, isoforms and PTMs
This is where the gel wins, and it is the reason the method survives. A 2D gel separates intact proteins, so a phosphorylation, a glycosylation, a truncation or an alternative splice product moves the spot. The modification is visible before you know what it is. Marcus, Lelong and Rabilloud describe 2D electrophoresis as the only currently available technique able to separate complete proteins over a wide pI and molecular mass range, and note that a series of histone modifications, including propionylation, butyrylation, malonylation, succinylation and crotonylation, were first found through the unsupervised view a 2D gel gives.
Shotgun proteomics sees modifications only on the peptides it happens to detect, and only a fraction of the tryptic peptides of any protein are detected. The bottom-up approach is, in Chait’s words quoted by the same review, suboptimal for determining modifications and alternative splice variants. Modification-specific enrichment can recover some of this, at the cost of a separate experiment per modification type.
In one comparison the review cites, a schizophrenia study found ten protein differences by 2D electrophoresis, and nine of them were missed by shotgun proteomics on the same material.
Quantitation
In 2D-DIGE, quantitation is built into the separation. Every spot on every gel is measured against a Cy2-labeled internal standard, fluorescence detection is linear over several orders of magnitude, and the statistics run on spot volumes. The weakness is co-migration: a spot is not always one protein, and where two proteins share a spot the measured change belongs to the pair. In roughly 30% of spots the dominant protein accounts for less than 70% of the signal.
In LC-MS/MS, quantitation is a choice of method. Label-free quantitation compares peptide signals across runs. Metabolic labeling (SILAC) or chemical labeling (iTRAQ, TMT) allows samples to be mixed and compared within a run; iTRAQ and TMT multiplex several samples with reporter ions and are accurate over about two orders of magnitude, with known interference from co-fragmentation. Each approach has its own assumptions, and quantitation is at the peptide level, inferred up to the protein.
Sample, cost, time and skill
2D-DIGE needs more sample than shotgun and more hands-on time per experiment: the review rates its sample consumption and time consumption both higher. It needs a fluorescence imager and the dyes, which are expensive, but no mass spectrometer until the identification step, and the identification can be outsourced spot by spot. The skill is in running reproducible gels and in image analysis.
LC-MS/MS needs less sample and less bench time per sample, but it needs a mass spectrometer, a chromatography system, the informatics to process the data, and people who can run all three. For a lab without those, it is an outsourced service with a per-sample price. Throughput, once the system is running, is far higher.
Limitations of each
Limitations of 2D-DIGE
Low-abundance proteins are poorly represented unless the sample is prefractionated. Very hydrophobic, very large, very small and extreme-pI proteins are under-represented. Co-migration confounds quantitation in a minority of spots. The method is labor-intensive and difficult to automate. Depth is limited to the high hundreds or low thousands of proteins.
Limitations of LC-MS/MS
Only a fraction of any protein’s peptides are detected, so proteoforms, splice variants and modifications are largely invisible unless specifically enriched. Protein abundance is inferred from peptides, and shared peptides make that inference ambiguous. Degradation products and truncations cannot be distinguished from the intact protein. There is no image to inspect, and no direct interface to biochemical follow-up techniques such as Western blotting. The instruments are expensive and the data processing is not trivial.
Complementarity: using both
The methods overlap less than people assume. In a maize chloroplast study comparing 2D electrophoresis, ICAT labeling and label-free LC-MS, only 20 proteins were quantified in common across 125, and the authors concluded that no single method provides qualitative and quantitative information on all the protein components of a complex mixture and that combining approaches yields complementary datasets.
The practical pattern is to use shotgun LC-MS/MS for depth and discovery across the whole proteome, and 2D-DIGE where the question is about proteoforms, modifications, degradation, or a targeted set of abundant proteins that must be quantified precisely, and wherever a visual, spot-level record is needed, as in host cell protein antibody coverage. In the words of the Proteomes review, the trade-off exchanges details and precision against speed and analysis depth, and 2D gel-based proteomics cannot be dismissed on simple fashion arguments.
Side-by-side comparison
| Property | 2D-DIGE | Shotgun LC-MS/MS |
|---|---|---|
| Separates | Intact proteins by pI and mass | Peptides by chromatography |
| Output | Spot map image | Protein list with abundances |
| Proteins per experiment | Roughly 800 gene products from 2,500 to 3,000 spots; more with large formats | Thousands |
| Proportion of cell protein mass seen | About 90% | Higher, includes low-abundance proteins |
| Proteoforms and PTMs | Seen directly as spot shifts | Seen only on detected peptides; enrichment needed |
| Degradation and truncation | Visible | Not distinguishable from intact protein |
| Quantitation | Internal standard, spot volume, linear over several orders of magnitude | Label-free, SILAC, iTRAQ or TMT; peptide level, inferred to protein |
| Co-migration or shared-peptide ambiguity | About 30% of spots have a dominant protein under 70% of signal | Shared peptides between proteins |
| Sample needed | More | Less |
| Time per experiment | More | Less |
| Instrument | Fluorescence imager plus MS for identification | LC-MS/MS system |
| Visual record | Yes | No |
| Best for | Proteoforms, PTMs, degradation, HCP coverage, targeted quantitation of abundant proteins | Global profiling, depth, discovery across the proteome |
Figures from references 1 and 2.
Analyzing the gel side
Whichever way the comparison falls for your experiment, the gel side is only as good as its image analysis. SameSpots aligns every DIGE image, detects one spot pattern across the whole experiment so that no spot is missing on any gel, normalizes to the internal standard, and runs the statistics. Identifications from the MS step are imported back onto the map, so the final report links each spot to its expression change and its identity.
Frequently asked questions
Q: Is 2D gel electrophoresis still used in proteomics?
A: Yes. It is no longer the method for global profiling, which shotgun LC-MS/MS does better, but it remains the only technique that separates intact proteins across a wide pI and mass range, which makes it the method of choice for proteoforms, modifications, degradation and assays that need a visual record such as HCP antibody coverage.
Q: What is the main advantage of 2D-DIGE over LC-MS/MS?
A: It separates intact proteins, so isoforms and post-translational modifications are visible as spot shifts without knowing in advance what to look for. Shotgun proteomics works on peptides and detects only a fraction of them.
Q: What is the main advantage of LC-MS/MS over 2D-DIGE?
A: Depth and throughput. A shotgun run identifies thousands of proteins, including low-abundance ones a gel cannot resolve, from less sample and less bench time.
Q: How many proteins can a 2D gel resolve compared with LC-MS/MS?
A: A typical 2D gel resolves 2,500 to 3,000 spots, about 800 gene products, which account for about 90% of cellular protein mass. Large-format gels have been reported at more than 10,000 spots. Shotgun LC-MS/MS identifies thousands of proteins in one run.
Q: What are the limitations of 2D gel electrophoresis compared with LC-MS?
A: Limited depth, poor representation of low-abundance and hydrophobic proteins, co-migration of proteins in some spots, and more sample and labor per experiment.
Q: Are 2D-DIGE and LC-MS/MS complementary?
A: Yes. In published comparisons the proteins quantified by each method overlap only partly, and combining them gives a more complete picture than either alone.
Q: What is the difference between top-down and bottom-up proteomics?
A: Top-down analyzes intact proteins; bottom-up digests proteins to peptides first. 2D gels are a top-down separation with bottom-up identification of the picked spots. Shotgun LC-MS/MS is bottom-up throughout.
References
1. Marcus K, Lelong C, Rabilloud T. What room for two-dimensional gel-based proteomics in a shotgun proteomics world? Proteomes. 2020;8(3):17. https://doi.org/10.3390/proteomes8030017
(Source for: 2,500 to 3,000 spots and about 800 gene products; about 90% of cellular protein mass; the dominant protein under 70% of signal in about 30% of spots; sample and time consumption ratings; 2D electrophoresis as the only technique separating complete proteins over a wide range; histone modifications first found by 2D gels; the Chait quotation on bottom-up limitations; the schizophrenia study; fluorescence linearity; the “fashion arguments” conclusion.)
2. Abdallah C, Dumas-Gaudot E, Renaut J, Sergeant K. Gel-based and gel-free quantitative proteomics approaches at a glance. International Journal of Plant Genomics. 2012;2012:494572. https://doi.org/10.1155/2012/494572
(Source for: more than 10,000 spots corresponding to more than 1,000 proteins; 2-DE limitations; iTRAQ and TMT multiplexing and accuracy over about two orders of magnitude; the maize chloroplast comparison with 20 of 125 proteins in common; the conclusion that no single method covers all components.)
3. Görg A, Weiss W, Dunn MJ. Current two-dimensional electrophoresis technology for proteomics. Proteomics. 2004;4(12):3665-85. https://doi.org/10.1002/pmic.200401031
(Source for: the contrast between an intact-protein map and peptide-based LC-MS/MS.)
Get the most from the gel side
If your question needs intact proteins, the analysis needs to be as precise as the separation. SameSpots gives every spot a value on every gel and runs the statistics on a complete dataset. Request a trial and analyze your own DIGE images.