Why Tissue Composition Fractions Need a Joint Check
Water, lipid, and protein fractions describe important features of a tissue, but they do not make sense as isolated numbers. A value can look plausible on its own and still be incompatible with the other reported fractions. For example, three fractions might each fall within a reasonable range yet sum to more than the tissue mass they are supposed to describe.
That inconsistency matters in physiologically based pharmacokinetic (PBPK) modeling because tissue composition can contribute to predicted drug distribution. A model that uses incompatible fractions may count too much tissue, assign an incorrect share of tissue to a drug-binding phase, or produce a misleading partition estimate. The first task is therefore to establish what each fraction means and whether the fractions can coexist.
Define the Measurement Basis Before Checking Values
The denominator is part of the definition. A water fraction calculated from wet tissue is not interchangeable with a water fraction calculated from dried tissue. On a wet-mass basis, the denominator includes tissue water. On a dry-mass basis, it does not. Similarly, a volume fraction describes a component's volume relative to a stated tissue volume; it is not a mass fraction.
Operational definitions also matter. “Lipid” might mean total extracted lipid, or it might refer to a particular lipid class. If total lipid already includes phospholipid, adding a separately reported phospholipid fraction counts some material twice. Protein and water estimates can also depend on assay definitions and sample processing. Before adding fractions, confirm that they refer to the same tissue, sampling conditions, denominator, and mutually exclusive components.
For nonoverlapping components reported on the same mass basis, the basic constraint is:
Here \(w_i\) is the mass fraction of component \(i\). If the listed components are exhaustive, the sum equals one. If the sum is less than one, the difference is an unreported residual that may include carbohydrates, minerals, nucleic acids, or other material. A residual is not an error by itself; it becomes a problem if a model assumes the listed fractions exhaust the tissue or if the residual is implausibly large for the intended definition.
There are no universal numerical ranges that apply to every tissue, species, age, or assay. Tissue-specific screening bounds should come from measurements or a justified reference source matched to the model context. Ranges used in the examples below are explicitly hypothetical teaching values, not reference physiology.
A Practical Audit for Bounds and Mass Closure
A useful audit separates two questions. First, is each reported fraction plausible for the particular tissue and measurement basis? Second, can the fractions coexist under mass closure? Passing one check does not imply passing the other. In particular, a set can pass every component-wise range check and still fail the joint sum constraint.
Identify whether values are wet-mass, dry-mass, or volume fractions, and whether the reported categories overlap.
Compare each value with tissue- and method-appropriate bounds. Treat a failed check as a reason to investigate, not as proof that the measurement is wrong.
For nonoverlapping mass fractions on one basis, calculate the sum and compare it with one. Record the residual if the sum is below one.
Check units, denominators, rounding, tissue identity, category overlap, and whether the inputs came from compatible samples. Do not silently rescale fractions to force closure.
Worked Example: Checking Fractions Against Illustrative Bounds
Suppose a hypothetical tissue dataset reports wet-mass water, total lipid, and protein fractions of \(0.78\), \(0.06\), and \(0.13\). For this example only, assume an analyst has chosen screening bands of \(0.65\)–\(0.85\) for water, \(0.02\)–\(0.12\) for lipid, and \(0.08\)–\(0.20\) for protein. Assume the three categories are nonoverlapping and measured on the same wet-mass basis.
Each fraction is within its stated illustrative band. The sum is:
The unreported residual is \(1 - 0.97 = 0.03\), or 3% of wet tissue mass. The set passes the component-wise checks and satisfies the joint constraint because \(0.97 \leq 1\). Whether a 3% residual is substantively plausible still depends on what the assays included and what other tissue components are expected; arithmetic closure alone does not establish biological validity.
Why Incompatible Fractions Distort Distribution Predictions
A simple composition-based calculation illustrates the consequence of mass closure. Suppose each tissue component has a drug-storage capacity \(q_i\), defined here as the amount of drug associated with that component per unit reference concentration, per unit mass of component. Its units are volume per mass, such as L/kg. Under an additive approximation, the total capacity per mass of tissue is the sum of each component's capacity weighted by its tissue mass fraction.
This expression assumes the fractions use the same mass basis, the components do not overlap, and their contributions add at equilibrium at a common reference concentration. It is a teaching representation of composition-weighted storage, not a complete PBPK tissue-partition model. Real models can use other definitions and account for additional biology; the equation is useful here because it shows why a fraction error can affect a distribution estimate.
The result follows directly from adding component amounts. If the reference concentration is \(C_{\text{ref}}\), component \(i\) has mass \(m_i\), and its drug amount is \(m_i q_i C_{\text{ref}}\), then total amount is the sum over components. Dividing by total tissue mass \(m_{\text{tissue}}\) and by \(C_{\text{ref}}\) gives:
If an input set sums above one, using it directly assigns more than the full tissue mass to its components. The calculated capacity can then be inflated, especially when an overcounted component has a large \(q_i\). Dividing all fractions by their sum is not a general repair: it changes the reported measurements and assumes, without evidence, that proportional rescaling is appropriate.
Worked Example: How an Overcounted Lipid Fraction Changes Capacity
Use the valid hypothetical fractions from the previous example: water \(0.78\), lipid \(0.06\), protein \(0.13\), and residual \(0.03\). Suppose the illustrative component capacities are \(q_{\text{water}}=0.10\), \(q_{\text{lipid}}=20\), \(q_{\text{protein}}=2.0\), and \(q_{\text{residual}}=0.20\) L/kg of component. These values are hypothetical and are chosen to show the arithmetic, not to represent a particular drug or tissue.
The additive tissue capacity is:
Now suppose an analyst mistakenly uses lipid \(0.10\) and protein \(0.16\) while leaving water at \(0.78\). These reported components sum to \(1.04\), before any residual is added. If the analyst nevertheless applies the same weighted calculation to those three numbers, the result is:
That value is not a defensible estimate from a mass-closed tissue composition: the inputs assign 104% of the tissue mass before accounting for any other material. The larger result is driven mostly by the extra lipid fraction because its assumed component capacity is high. The correct response is to investigate or replace the incompatible inputs, not to interpret the 2.398 L/kg result as evidence of greater tissue distribution.
Basis Conversion and Common Failure Modes
When fractions are reported on different bases, do not combine them directly. For instance, if 100 g of wet tissue contains 72 g of water, 8 g of lipid, 15 g of protein, and 5 g of other material, the wet-mass fractions are \(0.72\), \(0.08\), \(0.15\), and \(0.05\). The dry mass is \(100-72=28\) g, so the dry-mass lipid fraction is \(8/28=0.286\), and the dry-mass protein fraction is \(15/28=0.536\). These dry-mass values cannot be added to the wet-mass water fraction \(0.72\) as if they shared a denominator.
Worked Example: Detecting a Mixed-Basis Sum
An analyst combines the wet-mass water fraction \(0.72\) from the sample above with dry-mass lipid and protein fractions \(0.286\) and \(0.536\). The apparent sum is:
This sum does not show that the sample contains 154.2% of its mass in those components. It is invalid because the water fraction is divided by 100 g wet tissue, while lipid and protein fractions are divided by 28 g dry tissue. On the common wet-mass basis, the relevant sum is \(0.72+0.08+0.15=0.95\), leaving \(0.05\) for other material. The basis must be corrected before closure is assessed or the values enter a model.
- Assuming every component must sum to one. The sum equals one only when the listed categories are exhaustive and nonoverlapping. Otherwise, an unreported residual is expected.
- Checking only individual ranges. Fractions can each look reasonable while their sum exceeds one. Check both individual bounds and the joint constraint.
- Adding overlapping categories. A lipid subtype may already be included in total lipid. Confirm definitions before summing.
- Normalizing incompatible values without justification. Rescaling may conceal a wrong denominator, overlap, or measurement problem, and it changes the original inputs.
- Treating a screening bound as universal. A tissue- and method-specific reference is needed; a generic range can flag a question but cannot establish plausibility for every population.
Interpreting an Audit in a PBPK Workflow
A failed check should trigger a traceable review. Start with the source and record the tissue, species or population, sample condition, assay definition, units, and denominator. Then check arithmetic and rounding, followed by possible overlap and missing residual categories. If the discrepancy remains, seek compatible measurements or document why an input is retained despite the concern.
Composition checks complement checks on organ volumes, blood flows, and other physiological inputs: each constrains a different part of the model. They do not by themselves validate a tissue partition prediction. In particular, passing mass closure does not prove that the chosen component capacities are appropriate, that the tissue is adequately represented, or that the predicted distribution is correct. Conversely, a predicted value that appears reasonable does not excuse a composition set that violates its stated basis.
Check Your Understanding
Answer each question, then compare with the brief solutions below.
- Three nonoverlapping wet-mass fractions are water \(0.70\), lipid \(0.08\), and protein \(0.17\). What is the residual, and does the set satisfy mass closure?
- Why can a dry-mass lipid fraction not be added directly to a wet-mass water fraction?
- A tissue has fractions \(0.75\), \(0.12\), and \(0.18\), each within its own screening band. What joint check is still required?
- In the additive capacity expression, which fraction change would have the larger effect: increasing a component with \(q_i=0.2\) L/kg by 0.01, or increasing one with \(q_j=10\) L/kg by 0.01? Explain.
Solutions. (1) The sum is \(0.70+0.08+0.17=0.95\), so the residual is \(1-0.95=0.05\), or 5%; the set satisfies the constraint because the sum is below one. (2) The denominators differ: wet mass includes water, while dry mass excludes it, so the fractions are not directly comparable or additive. (3) Calculate the sum: \(0.75+0.12+0.18=1.05\). Despite passing separate bands, the fractions violate mass closure if they are nonoverlapping and on one basis. (4) The change in capacity is \(\Delta Q=\Delta w\,q\). The first change contributes \(0.01(0.2)=0.002\) L/kg tissue; the second contributes \(0.01(10)=0.10\) L/kg tissue, so increasing the fraction with \(q_j=10\) has the larger effect under the stated additive assumptions.