Soil carbon-13 DNA stable isotope probing, often shortened to 13C DNA-SIP, is used to investigate which microbial populations incorporated carbon derived from a labelled substrate into newly synthesised DNA. If you are trying to read a soil DNA-SIP graph, heavy versus light DNA fractions, a caesium chloride density gradient or 16S amplicon profiles from labelled fractions, begin with the substrate supplied and the unlabelled control. Heavy DNA is not an organism name, and a high-density fraction is not automatically proof of carbon assimilation.
The essential comparison is whether a particular sequence shifts in buoyant-density distribution after exposure to 13C-labelled material compared with a matched 12C control. DNA density also reflects its natural base composition, sample preparation and fractionation. That is why a heavy fraction without an appropriate control cannot identify which organisms consumed the tracer. DNA-SIP records isotope incorporation in a genetic molecule under stated conditions, not a complete map of soil carbon flux.
This article explains soil 13C DNA-SIP interpretation from sampling and incubation through DNA extraction, CsCl gradient separation, density measurements, fraction sequencing, weighted-density analysis, isotope attribution and cautious ecological conclusions. Worked values are illustrative, not results from a particular laboratory experiment. The aim is practical: read the plot, understand what its fractions mean, check the controls, and express evidence about labelled taxa without describing every sequence as an active carbon consumer.
Quick answer: what does a heavy DNA fraction tell us?
During a DNA-SIP experiment, soil is incubated with material enriched in the stable isotope carbon-13. Microbes that assimilate labelled carbon into newly made DNA may produce DNA that migrates to a higher buoyant density during equilibrium centrifugation. Researchers collect fractions and examine how much DNA or how many target sequences appear at each measured density. A comparison with an unlabelled treatment helps distinguish isotope-induced shifts from naturally dense DNA.
The appropriate conclusion is conditional: a particular marker sequence showed evidence of enrichment in DNA fractions whose density shifted relative to the unlabelled control. That suggests incorporation into newly synthesised nucleic acid under the experiment. It does not necessarily demonstrate direct uptake of the originally added molecule. Carbon can move through the soil food web or pass through intermediates before another population incorporates it.
Start with the carbon source and the research question
A labelled substrate might be glucose, cellulose, xylose, a plant-derived mixture or another carbon source chosen to probe a particular metabolic route. The ecological question should match that source. A 13C-glucose study investigates incorporation of carbon from glucose or its downstream products; it does not automatically measure decomposition of every component of soil organic matter. A labelled cellulose incubation answers a different question about complex substrate utilisation and carbon transfer.
Before reading the figure, write a one-sentence question: which taxa incorporated carbon from the administered material during the stated incubation? Next identify the amount of carbon added, its isotopic enrichment and whether the dose perturbed the system. A very large artificial carbon pulse can select for fast-growing opportunists that would not dominate under natural concentrations. SIP helps identify potential carbon processors within the experimental conditions; generalisation to entire ecosystems needs an additional argument.
Why stable carbon-13 works as a tracer
Carbon occurs naturally as several isotopes, including carbon-12 and carbon-13. The two are chemically similar but differ in nuclear mass. Carbon-13 is stable, unlike radioactive carbon-14, and its enrichment can be measured analytically. When a microorganism incorporates labelled carbon into newly synthesised nucleotides, the DNA’s mass and buoyant-density behaviour can change. The tracer therefore connects a substrate input to a recoverable biological macromolecule.
That connection must be interpreted carefully. The isotope reveals a path of incorporated atoms, not a visible history of every metabolic intermediate. Label can be respired to CO2, used to synthesise many compounds, transferred among organisms or diluted by unlabelled carbon pools. The appearance of 13C in DNA demonstrates incorporation under the model but does not by itself reveal which biochemical enzyme catalysed the initial transformation.
Sampling and biological replication still come first
Soil microbial communities are heterogeneous at centimetre and sometimes smaller scales. One gram may differ from neighbouring material in root density, moisture, oxygen and available substrate. Collection depth, mixing, cold storage, sieving and preincubation can alter the starting community. Replication across independent soil units is therefore central to deciding whether a labelled signal is robust rather than a peculiarity of one microcosm.
A sample split into several tubes after homogenisation creates useful technical or incubation replicates, but these do not necessarily represent independent field sites. Ask how many independently collected soils or experimental units were analysed, and whether controls and labelled treatments were randomised across preparation batches. A plot with many gradient fractions from one soil is not equivalent to a study with many independent soil observations.
Define the unlabelled comparison precisely
The strongest basic control is usually an otherwise comparable incubation supplied with the same substrate in its natural isotopic composition. The aim is to expose microbes to similar chemical conditions while changing the isotope label. A no-addition control can answer another question, but it is not always interchangeable with an unlabelled-substrate comparison because substrate addition itself can reshape growth and community composition.
Check whether the labelled and unlabelled treatments used the same amounts, durations, soils and extraction procedures. If their chemical conditions differ, a density or community shift may reflect treatment differences beyond the isotope. Where possible, include time-zero material and extraction blanks. A clear report says which comparison supports the conclusion rather than calling any untreated tube a perfect control.
Growth and DNA synthesis are part of the signal
Conventional DNA-SIP generally relies on labelled carbon reaching DNA through synthesis or turnover. A microbial cell may metabolise a labelled substrate without synthesising enough DNA during the incubation to show a clear density shift. Conversely, a population may divide using labelled intermediates produced by other organisms. Thus DNA-SIP detects a particular part of microbial activity, not all respiration or every enzyme reaction.
This distinction becomes important in short incubations. Organisms can oxidise carbon for energy or maintain cellular processes without producing a readily detectable labelled DNA fraction. Absence of a heavy-DNA signal therefore does not prove inability to utilise the substrate. A carefully written statement concerns detectable DNA incorporation within the timing and sensitivity of the experiment, not an absolute boundary on physiological capabilities.
Carbon cross-feeding can change the interpretation
Imagine a primary decomposer breaking labelled cellulose into soluble compounds. A neighbouring organism may consume those compounds or products of the primary decomposer and incorporate carbon-13 into its DNA. Both populations can appear labelled, even though only one initially attacked the cellulose polymer. Additional recycling through CO2 fixation or microbial residues can also transfer the tracer through a community.
A late heavy-DNA signal is therefore not always evidence of direct utilisation of the original material. Compare sampling times, substrate chemistry and known metabolic capabilities. Shorter pulses may sometimes reduce secondary transfer, but do not eliminate every interaction and may miss slow-growing populations. Labelled taxa are participants in carbon incorporation pathways under the experiment; proving direct substrate attack requires more targeted evidence.
Soil preparation affects the available carbon pool
Water content, aeration, aggregation and nutrient availability regulate how added substrate moves and which microorganisms gain access. Adding labelled solute to an intact core differs from mixing it uniformly into a slurry. A homogenised treatment may disturb microhabitats and expose previously protected carbon. Temperature and incubation vessel geometry can also influence gas exchange and microbial growth.
Read the laboratory protocol before assuming every unit experienced the same exposure. If carbon concentration was set by solution volume, check whether soil dry mass and water content were accounted for. Compare isotopic enrichment on the same atom-fraction basis. A sound interpretation states what was done to the soil and how that procedure may differ from undisturbed field conditions.
DNA extraction creates another selective filter
Different microbial cells release DNA with different efficiencies. Mineral surfaces bind nucleic acids, and humic compounds can interfere with purification. A labelled organism may be underrepresented because it resisted lysis or its DNA was lost during cleanup. Differences in extraction batches between labelled and control treatments can alter apparent representation even if the biological sample remains similar.
Readers should locate extraction mass, soil dry-weight conversion where relevant, bead-beating or other lysis details, purification steps and yield measures. Extraction blanks help identify reagent-associated DNA. A total-DNA yield is not necessarily a total living-cell count, because extracellular DNA and DNA from inactive cells may be present. DNA-SIP narrows inference to recovered molecules, so the extraction boundary should remain visible throughout the article.
Caesium chloride gradients separate by buoyant density
A common DNA-SIP method suspends DNA in caesium chloride solution and subjects it to prolonged ultracentrifugation until molecules distribute according to buoyant density. DNA richer in heavy isotopes may migrate differently, but naturally higher guanine–cytosine content can also increase density. The gradient is subsequently collected into successive fractions, each associated with a density interval or measured mean density.
The word ‘heavy’ is relative and method-specific. It does not denote a single universal density value applicable to every sample or laboratory. The centrifugation setup, initial solution, rotor, handling and fraction collection contribute to what is recovered. A careful caption gives actual density units, normally grams per millilitre, rather than merely numbering fractions and expecting the reader to infer direction.
Fraction numbers do not necessarily run from light to heavy
Some labs collect fractions from the bottom of a tube, others from the top or by different recovery apparatus. Consequently fraction 1 may be among the densest or lightest depending on the collection convention. Reversing the ordering can invert a biological story. Before translating any fraction chart, inspect the density axis or the sampling procedure itself.
If the plot labels fractions but not their densities, do not guess which direction means heavier DNA. Ask for the fraction-density measurements or the method describing collection order. This is a straightforward example of why technical terms and measurement directions matter more than a beautiful figure. The correct sentence uses measured buoyant density, not an assumed ordinal label.
A density profile is a distribution, not one point
Recovered DNA spans several fractions because a gradient has finite resolution and the extract contains diverse sequences and molecular populations. A marker can appear across both light and heavy fractions, with one or more peaks. An isotope-associated shift may alter the distribution without producing a perfectly separate new peak. The shape depends on incorporation, base composition, extraction and measurement noise.
Avoid declaring that every read found above a chosen density threshold belongs to an active organism. The threshold is an operational comparison that needs controls. Inspect the entire distribution and whether quantitative marker counts were measured across fractions. A robust interpretation often uses the contrast of full profiles rather than the presence of one apparently impressive bar.
Natural GC content can imitate heavy DNA
DNA with a high proportion of guanine and cytosine bases can have greater buoyant density than DNA with lower GC content, even without carbon-13 labelling. Different microbial lineages can therefore occupy different parts of the gradient in an unlabelled control. This creates a central false-positive risk: treating naturally dense sequences as if the tracer made them heavy.
Researchers address this by analysing matched unlabelled profiles and, in quantitative approaches, shifts in a sequence’s density distribution. An organism present in dense fractions in both treatments may not have incorporated detectable label. The critical question is what changed relative to its own control distribution, not whether it happened to sit on the ‘heavy’ side of the tube.
How to interpret a control-normalised density shift
Imagine an illustrative ASV with a weighted mean buoyant density of 1.700 grams per millilitre in the unlabelled control and 1.712 grams per millilitre in the labelled treatment. The observed shift is positive 0.012 grams per millilitre. This difference is a useful summary of the two distributions if the means were calculated consistently and enough reliable reads or quantitative measurements supported them.
It would be incorrect to convert 0.012 directly into a universal percentage of carbon assimilated without a calibrated isotope-density model and appropriate assumptions. Taxon-specific abundance, baseline GC density and fractional recovery may influence the estimate. State the density shift with its units, replicate uncertainty and control definition first. Then describe isotope incorporation as supported by the method rather than inventing a precise uptake rate.
A weighted mean is sensitive to fraction coverage
Suppose an illustrative sequence is recovered 70% in a fraction centred at 1.700 grams per millilitre and 30% at 1.724 grams per millilitre. Its two-fraction weighted mean is 0.70 times 1.700 plus 0.30 times 1.724, or 1.7072 grams per millilitre. The arithmetic is simple; the scientific challenge is knowing whether those weights truly reflect the sequence distribution.
If fractions at either end were lost, ignored or below detection, the weighted mean may shift artificially. Fractions of differing size require careful normalisation, and sequencing-read counts can be compositional. Check whether the denominator includes every collected fraction, which quantities were used as weights and whether DNA concentration or qPCR measurements were combined with relative read abundance.
Relative ASV fractions and absolute abundance are different
A heavy-fraction sequence profile might show an ASV representing 40% of reads. That does not mean 40% of total soil cells incorporated carbon-13. The fraction contains only a selected slice of recovered DNA, processed and sequenced with method-specific biases. Within that fraction the ASV shares a read denominator with other sequences, not with every microorganism in the original soil.
A full quantitative approach can measure target abundances through fractions, but its gene-copy assumptions remain important. When presenting a bar chart, write ‘40% of classified reads in the designated density fraction’ rather than ‘40% of active soil bacteria.’ This one phrase prevents the analysis from crossing an unjustified denominator boundary.
Sequencing heavy and light fractions requires care
After gradient collection, researchers may quantify DNA, amplify marker genes from chosen fractions and sequence them. Different fractions can contain different amounts of DNA, and PCR is sensitive to template input and amplification bias. A low-biomass fraction may generate noisy signals or fail to amplify a rare sequence. Pooled heavy fractions can simplify analysis while discarding information about the detailed density distribution.
Read which fractions were sequenced and how they were selected. If only two composite pools were analysed, the study may support a different resolution of inference from a project that measured every fraction quantitatively. A label such as ‘heavy fraction’ needs an explicit collection rule. Without that rule, the apparent difference may reflect researcher selection rather than a complete view of the gradient.
Heavy-fraction enrichment is not always taxon-specific incorporation
An ASV can seem more common in a heavy fraction because other taxa became less common there, even if its own absolute abundance changed little. This follows from the compositional nature of sequencing. Comparing only percentages can therefore suggest enrichment where the underlying gene copy distribution is more complicated. Contamination and variable fraction yields add further uncertainty.
The most informative comparison uses the density distribution of the particular feature between labelled and unlabelled conditions, ideally with quantitative measurements and independent replicates. Explain the normalisation method before making a statement about enrichment. Do not assume that a significantly higher read fraction in a selected pool proves direct isotope uptake by that lineage.
Conventional DNA-SIP versus quantitative SIP
Traditional DNA-SIP often seeks labelled organisms by examining the taxonomic content of denser DNA fractions after labelled substrate incubation. Quantitative stable isotope probing, or qSIP, uses a more explicit framework for assessing isotope-associated shifts across density fractions and estimating incorporation under a model. Implementations vary in tracer, fraction quantification and uncertainty treatment.
A different eduKateSG article explains oxygen-18 water qSIP, in which labelled water can mark newly synthesised DNA. That is not the same research question as tracing carbon from a 13C-enriched substrate. Water labelling and substrate carbon labelling probe different routes of incorporation, so their estimated enrichments should not be casually treated as interchangeable growth rates.
What a no-shift result can and cannot mean
A sequence with little apparent density shift might belong to a population that did not incorporate detectable tracer during incubation. It could also reflect slow growth, low initial abundance, insufficient label, rapid turnover, inadequate DNA yield, technical noise or a target outside the marker’s coverage. One cannot conclude that the organism is incapable of using the carbon source under all circumstances.
A useful negative finding specifies detection limits, sample depth, fraction resolution and experimental duration. Compare the control and labelled distributions to see whether the shift is genuinely absent or merely uncertain. When a taxon is of special interest, follow-up targeted measurements may be warranted. Good interpretation makes non-detection a bounded observation instead of turning it into a permanent biological identity.
Timing separates early assimilation from later recycling
Short incubation periods may preferentially reveal rapid incorporation, while longer incubations allow slower-growing taxa to synthesise DNA and permit secondary carbon transfer. The same species can appear at a different point in a time-course because the substrate pool, environmental conditions and interactions changed. A heavy-DNA profile at day two and day thirty therefore need not describe the same ecological event.
Plot density distributions by time, not only by treatment. Check whether substrate concentration or isotope enrichment changed as the experiment progressed. If labelled carbon entered microbial residues, predators or secondary metabolites, a later signal might reflect cross-feeding. A time series can clarify the sequence of events but rarely establishes direct substrate uptake by itself.
Carbon mass balance offers a different check
Labelled carbon added to a microcosm may be respired as carbon dioxide, retained in biomass, transformed into metabolites or remain in soil pools. DNA-SIP examines a relatively specific portion of incorporation. It does not account for every atom in the system, and labelled DNA may contain only a small share of all assimilated carbon.
Where researchers also measure labelled CO2, substrate disappearance or other pools, compare those quantities carefully. A strong heavy-DNA signal for one lineage does not determine the proportion of input carbon mineralised. Such questions require complementary mass-balance methods, with units expressed consistently as carbon mass or isotope atom fractions. Do not add proportions from unrelated analytical denominators and call the result total carbon recovery.
What does isotope atom fraction mean?
Atom fraction is the proportion of atoms of a given element that are the specified isotope. Natural carbon includes a relatively small background of carbon-13, so researchers often report enrichment above natural abundance or relative to an unlabelled control. Atom fraction, atom percent, atom percent excess and delta notation are related but not synonymous. A number lacking its isotope convention is ambiguous.
The distinction matters when interpreting a model that converts density shifts to incorporation. A shift of a few milligrams per millilitre is not itself an atom-percent increase. Model parameters, baseline density and nucleic-acid composition need to be defined. In student writing, keep measured density, isotope abundance and estimated carbon assimilation as separate columns. This simple table prevents one mathematical output from being presented as another.
Worked example: distinguishing DNA density from carbon share
Imagine a diagram showing two groups of DNA distributions. The unlabelled control peaks near 1.700 grams per millilitre for one feature. The labelled treatment has a broad distribution whose weighted mean is 1.711. The difference is 0.011 grams per millilitre. That observation supports a density shift under this example, but it does not imply ‘1.1% of soil carbon moved’ or that ’11 milligrams of carbon per litre’ was incorporated.
Those latter statements misuse the units. Grams per millilitre describes buoyant density of a solution-separated nucleic-acid fraction, not soil carbon concentration. To estimate a substrate-derived carbon fraction, researchers need a justified calibration and compositional model. Unit checking is one of the most effective ways for a reader to catch invented conclusions in isotope papers.
Controls for isotope-independent gradient changes
The fractionation and extraction procedures should be consistent across labelled and unlabelled microcosms. Differences in initial solution density, centrifugation, fraction size or measurement instrument can introduce apparent shifts. The proportion of DNA at each density also changes with taxonomic composition and DNA recovery. A labelled incubation may stimulate one population more than another even without considering isotope incorporation.
Consequently, changes in the overall community distribution and within-taxon label-associated shifts should be distinguished. A robust study may include technical standards, gradient-density measurements, blanks and independent replicates. Inspect whether all samples passed the same quality thresholds. Reproducibility is not obtained by a plotting package; it is built into sampling, control conditions and measurement.
When GC content and label incorporation overlap
A sequence from an organism with naturally high GC content may overlap the density range of strongly labelled DNA from a lower-GC organism. If the analysis simply separates the tube into one ‘heavy’ half and one ‘light’ half, the two could become confounded. Richer fraction-by-fraction information and control profiles help disentangle these effects.
This is one reason a single universal heavy-DNA cutoff is risky. Ask whether the authors report isotope incorporation based on a shift relative to controls or simply presence in dense fractions. If only presence is available, the conclusion should be restricted. The method remains useful, but readers deserve to know when the observed signal cannot unambiguously distinguish naturally dense DNA from isotope-enriched DNA.
RNA and PLFA tracing answer related but different questions
DNA-SIP typically requires label incorporation into DNA, often associated with cell synthesis or turnover. RNA-SIP can probe incorporation into RNA, potentially under different response times and turnover behaviour. Phospholipid fatty acid SIP measures label associated with particular extracted lipid biomarkers and can give broader group-associated evidence without the same genetic-sequence resolution.
The preceding guide to soil 13C PLFA-SIP explains fatty-acid enrichment and its own denominators. It should not be compared directly with DNA-SIP as though a lipid isotope ratio were a count of labelled ASVs. Different biomolecules turn over differently and carry different levels of taxonomic specificity. Combining them can sharpen a hypothesis only when their distinct measurement meanings remain clear.
DNA-SIP cannot establish precise metabolic pathways alone
If a sequence becomes heavy after incubation with labelled cellulose, the result is evidence consistent with carbon incorporation from that supplied source or products derived from it. It does not identify which cellulase the organism expressed, whether it was the first cell to cleave cellulose or the fraction of the community’s total carbon budget supplied by cellulose. Taxonomic identity and functional mechanism remain distinct questions.
Shotgun metagenomics, metatranscriptomics, metaproteomics and enzyme assays can contribute complementary evidence. The soil shotgun metagenomics guide distinguishes gene potential from observed activity. A persuasive ecological narrative follows measured carbon incorporation, predicted capabilities and independently observed reactions without blending them into one unsupported sentence.
Statistical evidence needs biological replication
A labelled versus unlabelled comparison made on one microcosm provides a useful observation but little information about variation among independent soils. Fraction sequencing can produce thousands of reads, but reads from one experiment are not thousands of biological replicates. Treating them as independent field samples would create false confidence.
For each target taxon, a responsible analysis considers variation in estimated density shifts across independently prepared biological units, fraction sampling and technical measurement uncertainty. Report effect sizes and uncertainty intervals where appropriate. A small shift consistently replicated may be more convincing than a dramatic difference seen in one tube. Where many taxa are evaluated, multiple-comparison issues can further complicate statistical labels.
The danger of selecting only attractive taxa
Large SIP feature tables contain many sequences. It is tempting to highlight whichever taxon shows the largest apparent heavy-fraction peak and construct a story around its known ecology. That is vulnerable to selection bias, rare-read noise and incorrect taxonomy. If the taxon belongs to a broad family with diverse metabolic behaviours, its name does not prove a specialised carbon-use function.
The solution is to state the original hypotheses and selection criteria, disclose the full taxon-level distribution and distinguish exploratory findings from confirmatory tests. It is fine to discover an unexpected candidate. Present it as a candidate supported by the measured isotope pattern, then explain which follow-up experiments could confirm direct assimilation or specific biochemical roles.
How to read a published density-gradient figure
Begin with the two axes: are they fraction number, buoyant density, DNA amount, gene copies or relative marker abundance? Next locate the unlabelled controls, then compare the same marker or taxon across isotope conditions. Check whether the ordinate sums across fractions or reflects a within-fraction proportion. Read time points, replicates and the definition of ‘heavy’.
Once those elements are clear, identify whether the evidence is a distribution shift, enrichment within a selected pool or a model-derived atom-fraction estimate. Do not leap to population growth or substrate-specific enzyme activity. A good description includes the exact measurement, a numerical contrast if reliable, and one key limitation. The diagram should lead the explanation rather than merely illustrate a conclusion chosen beforehand.
A second worked case: proportion is not uptake rate
Suppose a heavy DNA fraction contains 600 reads of ASV-A among 2,000 retained reads. Its share is 30%. Another time point yields 900 reads among 6,000, or 15%. An uncareful reader might say the organism’s carbon uptake halved. But the numbers are shares of sequence reads within two selected fractions with different total library sizes.
Neither 30% nor 15% describes a measured carbon flux per hour, and the changing fraction pool can affect the denominator. To study uptake rate, one needs calibrated isotope incorporation, suitable biomass or nucleic-acid quantities, sampling times and a quantitative model. The correct exercise answer is that ASV-A’s proportion of the selected fraction reads declined from 30% to 15%, with no standalone conclusion about its absolute substrate assimilation rate.
Common mistakes and targeted corrections
One mistake is equating naturally high-density DNA with 13C-labelled DNA. Correct it by comparing matched unlabelled gradients. Another is calling every enriched ASV a direct degrader of the administered substrate. Correct it by discussing intermediate metabolites and cross-feeding. A third is treating a percentage of reads in a heavy fraction as a percentage of soil cells or a carbon uptake rate. Correct it by preserving denominator and units.
Another mistake is treating one fraction as representative of the entire gradient. Correct it by inspecting full density distributions or acknowledging that they were not measured. A fifth is treating an absent shift as proof of no substrate utilisation. Correct it by considering slow growth, labelling time, detection sensitivity and alternative physiological routes. These repairs improve the inference without weakening the genuine experimental achievement.
Independent transfer exercise: a high-GC taxon
A taxon occurs in the densest fractions in both labelled and unlabelled samples, and the means of its density distributions are nearly identical. Does the evidence show that it assimilated the tracer? No. Its natural sequence composition may explain its density, and the experiment did not demonstrate a control-relative shift for that taxon. The appropriate conclusion is that 13C incorporation was not resolved by this comparison, within its measurement limits.
Now imagine a second taxon shifts consistently to higher density across biological replicates after labelled treatment, with identical fraction collection and a clear control distribution. The evidence for labelled DNA is stronger. Does that necessarily establish direct uptake of the original substrate? No; carbon may have passed through intermediates. The improved evidence concerns incorporation into DNA, not the entire metabolic route.
Questions readers frequently ask
Does a heavy DNA fraction always contain labelled organisms? No; naturally dense DNA can occupy it. Why use an unlabelled substrate control rather than no substrate? It helps separate isotope effects from chemical substrate effects. Is DNA-SIP a direct growth-rate measurement? Not automatically; synthesis and incorporation require interpretation. Can slow-growing microbes use label without a detected shift? Yes, depending on sensitivity and incubation timing.
Does 13C-DNA-SIP replace the 18O-water qSIP method? No; tracers answer different questions. Is a labelled ASV a species? No, marker resolution varies. Does an increase in heavy-fraction sequencing reads prove more organisms? No, the reads have a selected and compositional denominator. Can labelled carbon reach organisms through cross-feeding? Yes. Can DNA-SIP measure all carbon lost as CO2? No, it is only one part of a broader carbon mass balance.
Evidence and further reading
A soil DNA-SIP study investigating 13C-labelled cellulose and xylose describes the importance of comparing fraction distributions with unlabelled controls, including natural GC-content effects: Unearthing the Ecology of Soil Microorganisms Using a High-Resolution DNA-SIP Approach. A subsequent methodological study discusses how soil microbial DNA carbon-13 abundance affects identification. These sources illustrate concrete study designs, not universal cutoffs transferable without verification.
For connected learning, use soil 18O-water qSIP, soil PLFA-SIP, soil ASV tables and cross-linguistic transfer. Each page resolves a different point at which terms, units or methods can be confused.
Final principle: read the shift, not merely the heavy side
DNA-SIP is most informative when a well-designed labelled incubation is compared against appropriate unlabelled controls, density fractions are measured accurately and sequence-level evidence is reported with realistic uncertainty. The crucial change is label-associated density behaviour relative to the taxon’s baseline, not simply the presence of a familiar name inside a dense fraction.
A confident reader can say what carbon was supplied, which DNA markers shifted, how much support the controls provide and which interpretations remain unproven. That is genuine technical translation: preserving the measurement chain from a labelled carbon atom to a biological hypothesis, without pretending that the final chart tells the whole story of the soil.