Written and maintained by CASRAI Editorial Board
Last updated
Reader task: a bench scientist or graduate researcher with too little DNA for sequencing, genotyping or qPCR needs to know what whole genome amplification is and which family of methods fits the question.
Short answer
Whole genome amplification (WGA) copies all of a small DNA sample, such as DNA from one cell or a limited clinical specimen, to produce microgram quantities for downstream assays. The two main families are multiple displacement amplification (MDA), an isothermal method using phi29 polymerase, and PCR-based methods such as DOP-PCR. Independent comparisons show each has different biases.
Definition and scope
WGA is a pre-analytical step, not a measurement. It increases the amount of template so that applications such as next-generation sequencing or locus-specific qPCR have enough input; the output is amplified DNA, not a sequence. It is related to, but broader than, targeted nucleic acid amplification tests, which copy a defined target rather than the whole genome. WGA also differs from in-silico PCR, which is a computational prediction of primer binding.
How MDA and PCR-based WGA work
Multiple displacement amplification (MDA)
According to QIAGEN’s bench guide (read 2026-11-07), MDA uses phi29 DNA polymerase, a strand-displacing enzyme with 3’→5′ exonuclease proofreading activity, with random hexamer primers at a constant temperature. Because the polymerase displaces downstream strands as it extends, the products form branched networks. The vendor reports an average product length above 10 kb (range 2-100 kb) and states an error rate up to 1,000-fold lower than Taq-based methods; treat the latter as a vendor claim. QIAGEN also describes MDA as running at 30 °C and advises against 95 °C heat denaturation of the template, preferring alkaline denaturation. MDA is a form of isothermal amplification; the branched-product mechanism is related to rolling circle amplification.
PCR-based WGA (DOP-PCR and PEP)
Degenerate oligonucleotide-primed PCR (DOP-PCR) and primer extension preamplification (PEP) amplify with partially random primers under thermal cycling. QIAGEN’s bench guide states that these methods can produce secondary-structure artifacts, incomplete coverage, locus dropout and fragments shorter than 1 kb; this is a vendor characterisation, and the independent data below should be weighed alongside it.
MALBAC
MALBAC is a third approach used mainly for single cells; it appears in independent comparisons with MDA and DOP-PCR rather than in the vendor material above.
Methods compared
The most useful independent evidence read for this page is Hou et al., GigaScience 2015;4:37 (doi:10.1186/s13742-015-0068-3), which compared seven commercial kits representing DOP-PCR, MDA and MALBAC in single-cell resequencing. The figures below come from that one study and may differ with other kits, cell types and sequencing depths.
| Question | DOP-PCR | MDA | MALBAC |
|---|---|---|---|
| Genome recovered at high depth (abstract) | about 6% | about 84% | about 52% |
| Duplication | highest of the three | not highest | not highest |
| Copy-number detection in simulation (sensitivity / specificity) | 94.15% / 94.00% | 74.04% / 67.93% | 91.40% / 87.80% |
| Reproducibility | more reproducible than MDA | less reproducible | more reproducible than MDA |
| SNV detection | not the authors’ recommendation | comparable to MALBAC (one kit detected 83.42% of SNVs relative to bulk DNA) | comparable to MDA |
| Chimeric rearrangements | not highlighted | reported, mostly intra-chromosomal translocations (82.08%) | not highlighted |
The authors suggest DOP-PCR for copy-number and aneuploidy work, and MDA or MALBAC where both SNV and CNV detection are needed, with MALBAC preferred when reproducibility matters most.
Which question points to which method
- Copy-number or aneuploidy question: the Hou et al. simulation favoured DOP-PCR; MDA’s uneven amplification between cells reduces CNV accuracy.
- Single-nucleotide variant question: MDA and MALBAC performed comparably in the same study.
- Need for long products or high fidelity: the vendor describes MDA as producing longer products (over 10 kb) with a proofreading polymerase.
- Fragmented or fixed input: QIAGEN notes that FFPE DNA often runs to several hundred base pairs against about 2 kb for good-quality input, so more template may be needed, and no simple method measures cross-linking.
Where it matters for labs
Because WGA changes the representation of the genome, results should be interpreted against how the amplified DNA will be used. QIAGEN’s vendor test amplified 47 loci about 10,000-fold with at most a 6-fold difference in representation between loci, which illustrates that some unevenness is expected even from a well-performing method. When planning a sequencing run, depth versus breadth of coverage is the metric on which these methods differ most in the study above.
Common misunderstandings
- “Unbiased” amplification: a vendor may use the word, but the independent comparison shows measurable coverage, dropout and duplication differences between methods.
- WGA is sequencing: it only increases template quantity.
- Amplified DNA equals native DNA: allele dropout (a mean of 12.47% for one MDA kit in Hou et al.) and chimeras mean amplified DNA is not a faithful substitute for the original.
- One method suits every question: the evidence above is question-dependent.
FAQ
What is whole genome amplification?
It is a set of methods that copy the entire genome of a small DNA sample to yield enough material for downstream analysis, mainly MDA with phi29 polymerase and PCR-based approaches such as DOP-PCR.
Is MDA better than PCR-based WGA?
Neither is better for every purpose. In Hou et al. (2015), MDA recovered far more of the genome than DOP-PCR, while DOP-PCR was more accurate for copy-number detection in simulation.
Sources
- QIAGEN bench guide, Multiple displacement amplification (WGA), read 2026-11-07: qiagen.com (manufacturer).
- QIAGEN, Unbiased whole genome amplification, read 2026-11-07: qiagen.com (manufacturer).
- Hou Y et al. GigaScience 2015;4:37, read 2026-11-07: PMC4527218.
