How much DNA is needed for sequencing?
Input DNA by platform, library type, and sample quality
Matic Broz, PhD12 min
Most DNA sequencing libraries start from between 1 nanogram and 1 microgram of DNA. A targeted amplicon panel can run on 10 ng, an Illumina human whole-genome library is usually made from 100 to 500 ng, PacBio HiFi needs about 500 ng per SMRT Cell, and a Nanopore ligation library of long fragments uses 1 µg. Sanger sequencing sits apart, needing a few nanograms of a short PCR product or 150 to 300 ng of plasmid.
The library kit sets the figure, and three things move it: whether the protocol amplifies the DNA by PCR, how long the fragments must stay, and how pure and intact the sample is. A nanogram of human DNA holds only about 310 copies of the genome, so the methods that need least DNA are the ones that copy it most.
How much DNA does each sequencing method need?
Manufacturer protocols current in October 2026 span three orders of magnitude, from 1 ng for a bacterial genome on Illumina to 1 µg for a long-fragment Nanopore library. The amounts describe DNA entering library preparation, measured before any shearing or amplification, not the finished library loaded onto the instrument.
| Method | Example kit | Recommended DNA input | What else the protocol requires |
|---|---|---|---|
| Targeted amplicon panel | AmpliSeq for Illumina | 1 to 100 ng; 10 ng per primer pool recommended | Works with blood and FFPE tissue |
| Exome capture from FFPE tissue | Illumina FFPE DNA Prep with Exome 2.5 Enrichment | 40 ng; 50 ng for poor-quality samples | DNA fragmented to 100 to 350 bp; QC delta Cq of 4 or less |
| Short-read genome, microbial | Illumina DNA Prep | As little as 1 ng | Extra PCR cycles and separate normalization below 100 ng |
| Short-read genome, human | Illumina DNA Prep | 100 to 500 ng recommended minimum | Fluorometric quantification; A260/280 of 1.8 to 2.0 |
| PCR-free whole genome | Illumina DNA PCR-Free Prep | 25 to 300 ng | Blood, saliva, or dried blood spots; not FFPE |
| PCR-free whole genome, older kit | TruSeq DNA PCR-Free | 1 µg | Illumina DNA PCR-Free Prep matches its performance in less time |
| PacBio HiFi | SMRTbell prep kit 3.0 | 500 ng per SMRT Cell with SPRQ chemistry; 1 µg on Sequel II | At least 70% of DNA longer than 10 kb (GQN 7.0 or higher) |
| Nanopore ligation | Ligation Sequencing Kit V14 | 1 µg for fragments over 10 kb; 100 to 200 fmol for 1 to 10 kb | Pore occupancy falls below 100 ng |
| Nanopore rapid | Rapid Sequencing Kit V14 | 100 ng; 200 ng per sample with Rapid Barcoding | High-molecular-weight DNA |
| Sanger | BigDye Terminator v1.1 | 1 to 50 ng of PCR product; 150 to 300 ng of plasmid | One clean template per reaction |
Two patterns run through the table. Methods that amplify a small target, such as amplicon panels, need least, because PCR makes millions of copies of each region before sequencing. Methods that read long native molecules need most, because every step that shears, size-selects, or repairs the DNA loses material, and none of it is replaced by copying.
How much DNA is needed for Illumina and whole-genome sequencing?
For a human whole genome on Illumina, 100 to 500 ng is the standard recommendation when the library is amplified by PCR, and the current PCR-free kit accepts 25 to 300 ng. Illumina DNA Prep accepts 1 to 500 ng or more and allows as little as 1 ng for small genomes such as bacteria. Below 100 ng the protocol adds PCR cycles and asks for libraries to be quantified and normalized separately.[1]
Skipping PCR removes amplification bias, and Illumina reports more even coverage across GC-rich, AT-rich, and repetitive regions of the human genome. Illumina DNA PCR-Free Prep lists 25 to 300 ng of input; the older TruSeq DNA PCR-Free kit, still sold, lists 1 µg.[4][5] A whole-genome requirement therefore belongs to a specific kit and protocol revision, not to the method in general.
Targeted methods go lower because they sequence only a fraction of the genome. AmpliSeq for Illumina panels accept 1 to 100 ng and recommend 10 ng per primer pool, including DNA from formalin-fixed, paraffin-embedded (FFPE) tissue.[2] Exome capture, which enriches protein-coding regions by hybridization, sits in between: Illumina's FFPE exome workflow requires 40 ng, rising to 50 ng for poor-quality samples.[3]
Input mass is only one part of planning a run. Coverage, platform capacity, and analysis determine sequencing cost, while library preparation and batching affect sequencing time.
How many genome copies are in a nanogram of DNA?
A nanogram of human DNA contains about 310 copies of the haploid genome, the single set of 23 chromosomes found in an egg or sperm. Piovesan and colleagues calculated 6.41 picograms of DNA in a male diploid nucleus, which carries two sets, from the GRCh38 reference sequence.[9] A picogram is a thousandth of a nanogram, so the conversion takes two lines:
The same nanogram is the DNA of about 155 diploid cells. This arithmetic explains the low end of Table 1. In 10 ng, the recommended amount for an amplicon panel, we calculated about 3,100 haploid copies, so a mutation carried by 1% of them is present on roughly 31 template molecules. At 1 ng it would be on about 3, and no amount of sequencing depth can recover molecules that never entered the tube.
The calculation also sets the floor for single-cell work. One human cell holds about 6.4 pg of DNA, so single-cell genome sequencing must first copy the whole genome many times over before any standard library kit can use it. The forensic benchmark discussed below rests on the same conversion: 500 pg is about 80 diploid cells.[10]
Long-read protocols sometimes count molecules rather than mass, because the same mass of longer fragments contains fewer of them. Taking an average base pair as about 650 daltons, 1 µg of 10 kb fragments is roughly 150 femtomoles, inside the 100 to 200 fmol that Oxford Nanopore recommends for fragments of that length.[7]
How much DNA is needed for PacBio and Nanopore sequencing?
PacBio HiFi libraries need about 500 ng of genomic DNA per SMRT Cell with the current SPRQ chemistry, and Nanopore needs 1 µg for a ligation library of long fragments or 100 ng for its rapid kit. Both figures are higher than short-read inputs because the DNA must stay long through library preparation.
PacBio's August 2026 SMRTbell prep kit 3.0 checklist sets a minimum of 500 ng per acquisition with SPRQ chemistry and 1 µg per SMRT Cell on the older Sequel II systems. Multiplexed samples can share that mass. Average library recovery is about 15% of the input, so 500 ng of genomic DNA yields roughly 75 ng of finished library.[6]
Quality governs PacBio yield as much as mass does. The checklist recommends that at least 70% of the DNA be longer than 10 kb, equal to a genome quality number (GQN) of 7.0 or higher on an Agilent Femto Pulse, and for whole genomes prefers more than half the mass above 30 kb. Lower-quality samples can still be sequenced, but they need more input and give less yield.[6]
Oxford Nanopore's Ligation Sequencing Kit V14 recommends 200 fmol for fragments under 1 kb, 100 to 200 fmol for fragments of 1 to 10 kb, and 1 µg for libraries above 10 kb. As input falls below 100 ng, pore occupancy deteriorates markedly. The Rapid Sequencing Kit V14 uses 100 ng of high-molecular-weight DNA, and the Rapid Barcoding Kit V14 requires 200 ng for each barcoded sample.[7]
Nanopore read length reflects the fragments supplied, so handling matters. Harsh pipetting, vortexing, repeated freeze-thaw cycles, and degraded extraction shorten reads even when the tube holds the full microgram.[7]
How much DNA is needed for Sanger sequencing?
A Sanger reaction needs 1 to 50 ng of a purified PCR product, depending on its length, or 150 to 300 ng of double-stranded DNA such as a plasmid. Thermo Fisher's BigDye Terminator v1.1 guide, revised in July 2026, recommends 1 to 3 ng for products of 100 to 200 bp, 3 to 10 ng for 200 to 500 bp, 5 to 20 ng for 500 to 1,000 bp, 10 to 40 ng for 1,000 to 2,000 bp, and 20 to 50 ng above 2,000 bp.[8]
The same guide recommends 25 to 50 ng of single-stranded DNA, 0.5 to 1 µg of cosmid or BAC DNA, and 2 to 3 µg of bacterial genomic DNA, the last because a primer's single binding site is diluted across a whole genome. Longer PCR products need more mass for the same reason: what matters is the number of primer binding sites, and a longer molecule weighs more per site.[8]
Sanger sequencing reads one template population per reaction, so a clean template matters as much as mass. Mixed PCR products or plasmids with more than one insert produce overlapping peaks even when the submitted quantity is correct. Primer design matters too, and Primer3 can design sequencing and PCR primers against a known target sequence.
Why DNA quality matters as much as quantity
A sample can meet the mass requirement and still fail if the DNA is impure, broken, or measured with the wrong assay. Most protocols in Table 1 set conditions on all three.
Purity is read from UV absorbance ratios. Illumina DNA Prep is optimized for an A260/280 ratio of 1.8 to 2.0 and an A260/230 ratio of 2.0 to 2.2; Oxford Nanopore gives about 1.8 as the A260/280 of pure DNA and notes that a low A260/230 can indicate phenol, guanidine, or carbohydrate carried over from extraction.[1][7] These contaminants inhibit the enzymes that build the library.
Concentration should come from a fluorometric assay such as Qubit, not from UV absorbance on an instrument such as NanoDrop. Absorbance counts RNA, free nucleotides, and single-stranded DNA along with double-stranded DNA, so it can overstate the usable mass; Illumina, PacBio, and Oxford Nanopore all specify dye-based quantification.[1][6][7]
Integrity matters most where fragments must be long. PacBio's GQN threshold and Oxford Nanopore's advice to size molecules over 10 kb by pulsed-field electrophoresis or Femto Pulse, rather than by Bioanalyzer, both follow from that.[6][7] Short-read FFPE workflows check integrity differently, with a qPCR quality test that flags samples whose fixed, degraded DNA would give poor coverage.[3]
How much DNA is needed for PCR, gels, cloning, and forensic identification?
A typical 50 µL PCR uses 5 to 50 ng of genomic DNA or 0.1 to 1 ng of plasmid DNA; an agarose gel can detect about 10 ng in a band; standard cloning often starts with 50 ng of vector; and forensic STR profiling has an optimum PCR input near 500 pg.
| Application | Published DNA amount |
|---|---|
| PCR | 5 to 50 ng genomic DNA or 0.1 to 1 ng plasmid DNA |
| Agarose gel | About 10 ng detection limit; up to 100 ng for a sharp band |
| Two-fragment cloning | 50 ng vector with a 3:1 insert-to-vector molar ratio |
| Forensic STR profile | About 500 pg optimum input, roughly 80 diploid cells |
PCR needs more genomic DNA than plasmid DNA because the target occupies a much smaller share of the molecules: a single-copy human gene appears once in each 3.2 pg haploid genome, while a plasmid is almost all target. Too much template raises the risk of nonspecific amplification, while too little reduces yield.[11]
For cloning, the insert amount depends on fragment length. A standard NEB ligation uses a 4 kb vector at 50 ng and a 1 kb insert at 37.5 ng, producing a 3:1 insert-to-vector molar ratio.[13] Copying "50 ng" without recalculating the insert gives the wrong number of molecules, the same mass-versus-molecules problem that Nanopore protocols solve by counting femtomoles.
For forensic identification, 500 pg is the optimum DNA mass entering STR amplification, not the amount that must be collected from an object. Swabbing, transfer, extraction, and purification can lose 20% to 80% of the DNA, so a practical sample needs more biological material than the final PCR input suggests.[10]
Across all of these uses, the useful question is how many intact copies of the target reach the reaction. Mass is the convenient proxy. It works when the genome size, fragment length, and purity are known, and it misleads when any of them is not.
