What is Dr. Zubair Khalid's research focus?

Dr. Zubair Khalid specializes in molecular virology, mRNA vaccine development, and computational biology, with a focus on avian pathogens like IBDV and Avian Reovirus.

Where is Dr. Zubair Khalid currently working?

Dr. Zubair Khalid is a Postdoctoral Research Associate at the University of Maryland (UMD), specifically within the Department of Animal and Avian Sciences.

Stability Definition: Meaning in Science and Biology

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-09-19

Stability is the tendency of a system to remain in, or return to, a defined state when conditions change. In biology, that "defined state" can be a folded protein, a population at carrying capacity, or a drug concentration inside a sealed vial, and each one is measured with completely different tools and units.

That context dependence is the single most important thing to understand about the word. A protein can be thermodynamically stable at 25 degrees C and still be degraded within minutes by a cell. A forest can be stable for a century and then flip to grassland after one drought. A tablet can hold 98 percent of its labeled potency for two years and still fail a dissolution test. Stability is not one property. It is a family of properties, each defined by what is being held constant, what is being disturbed, and how long the system is observed.

This guide defines stability across three contexts that students and researchers encounter most often: protein chemistry, ecology, and drug storage. It gives the key metric and unit for each, shows how stability is actually measured, and flags the misconceptions that cause the most errors on exams and in the lab.

What Stability Means in Science

The general definition of stability is the capacity of a system to resist displacement from a reference state, or to recover that state after a disturbance. Two ideas hide inside that sentence. Resistance is how much a system moves when pushed. Resilience is how quickly it comes back. A system can be high in one and low in the other, which is why ecologists rarely use the word "stable" without qualifying which kind they mean.

Physicists and chemists add a third layer: stability is always relative to a reference state and a timescale. A diamond is thermodynamically less stable than graphite at room temperature, yet the conversion takes longer than the age of the universe, so for practical purposes the diamond is stable. The reference state and the observation window define the answer.

Stability Is a Relationship, Not a Substance Property

Nothing is stable in the abstract. A protein is stable relative to its unfolded state under specific buffer conditions. A drug is stable relative to its labeled potency under specific temperature and humidity. A microbial community is stable relative to a baseline composition over a defined sampling interval. Change the reference, the conditions, or the window and the number changes.

This is why a "stability def" that works for one field fails in another. The sections below give the working definition, the metric, and a concrete example for each domain.

Stability in Protein Chemistry

In protein chemistry, stability is the free energy difference between the folded, native state and the unfolded, denatured state. A protein is stable when the folded state sits at a lower free energy than the unfolded state under physiological conditions.

Thermodynamic Stability and Delta-G

The standard measure is the Gibbs free energy of unfolding, written as delta-G of unfolding (ΔG_unfold) and reported in kilocalories per mole (kcal/mol). A positive ΔG means the folded state is favored. A negative ΔG means the protein would rather unfold.

Typical globular proteins have a ΔG of unfolding in the range of roughly 5 to 15 kcal/mol at 25 degrees C. That is a small number. It is equivalent to only a few hydrogen bonds or one or two buried hydrophobic side chains, which is why proteins are described as marginally stable. The margin is thin enough that a single point mutation can tip a protein from folded to aggregated.

A critical technical point: ΔG of unfolding cannot be measured directly under physiological conditions. As Ahmad explains, the value at zero denaturant concentration, written ΔG_D^0, can be neither measured nor predicted under physiological conditions [1]. What researchers actually do is add a strong chemical denaturant such as guanidinium chloride or urea, which destroys the noncovalent interactions holding the folded structure together, and then measure the transition as denaturant concentration rises. The value at zero denaturant is obtained by extrapolating the linear plot of ΔG versus denaturant concentration back to the y-axis [1]. That extrapolation is a model-dependent estimate, not a direct reading.

Folding, Denaturation, and Tm

Protein folding is the process by which a polypeptide chain adopts its functional three-dimensional shape. Denaturation is the loss of that shape without breaking the peptide backbone. Denaturation can be triggered by heat, extreme pH, chaotropic salts, or mechanical shear.

Thermal stability is usually reported as the melting temperature, Tm, in degrees Celsius. Tm is the temperature at which half the protein molecules are unfolded and half remain folded. A higher Tm means the protein tolerates more heat before losing structure. Tm is measured by techniques such as circular dichroism, differential scanning calorimetry, or fluorescence spectroscopy, each tracking a different signal that changes as the protein unfolds.

Tm and ΔG are related but not interchangeable. Tm is a temperature. ΔG is an energy. A protein can have a high Tm and a modest ΔG, or the reverse, depending on the shape of the unfolding curve.

Measuring and Engineering Protein Stability

Two practical approaches dominate. The first is experimental: titrate the protein with denaturant, record the unfolding curve, and fit it to a two-state or three-state model. The second is computational. Modern inverse folding models treat protein stability as a probabilistic property and use sequence design algorithms to propose variants with improved thermostability. Stern and colleagues introduced probabilistic definitions of protein stability and conformational specificity and showed that their BayesDesign algorithm increased the thermostability of the NanoLuc enzyme and increased the conformational specificity of the WW structural motif [2].

Stability in a test tube is not the same as stability inside a cell. Cells actively destroy proteins through the proteasome and lysosome, so a protein with excellent thermodynamic stability can still have a short half-life if it carries a degradation signal. Nightingale and colleagues used multiplexed proteomic and transcriptomic screens to quantify active protein degradation during human cytomegalovirus infection and identified 35 proteins enriched in antiviral restriction factors that were degraded during early infection [3]. One example is helicase-like transcription factor (HLTF), a DNA helicase involved in DNA repair that potently inhibits early viral gene expression but is rapidly degraded during infection. The viral protein UL145 facilitates that degradation by recruiting the Cullin4 E3 ligase complex [3]. This is a clean illustration of the distinction: thermodynamic stability describes folding, while biological stability describes persistence in a living system.

Stability in Ecology

In ecology, stability describes how a population or community responds to disturbance. Disturbance means any event that removes biomass or changes conditions, such as fire, drought, disease, harvesting, or a flood.

Resistance and Resilience

Ecologists split stability into two measurable components.

Resistance is the magnitude of change caused by a disturbance. A highly resistant community barely shifts in composition or abundance when the disturbance hits.

Resilience is the speed of return to the pre-disturbance state. A highly resilient community may change a lot but recovers quickly.

These two properties can trade off. A mature forest may resist a drought well because deep roots buffer water stress, but if it does cross a threshold and die back, recovery can take decades. A grassland may be easily knocked back by drought but regrow within a single season. Both are stable in different senses.

Populations, Communities, and Persistence

At the population level, stability often means persistence: the probability that a population survives over a defined time horizon. At the community level, stability usually means constancy of composition, or the absence of large swings in relative abundance among species.

Stochastic population dynamics models make these ideas quantitative. Ponciano and colleagues fitted stochastic models to multi-species bacterial time series from the human vaginal microbiome, explicitly accounting for three sources of variability: stochastic biotic and abiotic forces, ecological feedback, and sampling error [4]. Using estimates of interaction strengths, they built a Risk Prediction Monitoring tool that estimates how persistence probabilities for a bacterial group of interest change over time, mirroring extinction-risk tools used in conservation biology [4]. The key insight is that community stability depends on the nature and intensity of interactions within and between species, not just on external conditions.

A related concept applies to host-associated microbial ecosystems. The intestinal microecology is described as having characteristics of stability and dynamic balance, and disruption of that balance is linked to a range of diseases [5]. Stability here is not stasis. It is a moving equilibrium maintained by feedback between the host and its microbes.

Measuring Ecological Stability

Common measurements include:

Coefficient of variation of population size over time, which captures variability.
Return time after a perturbation, which captures resilience.
Intraclass correlation coefficients (ICC) to assess whether a measured behavior or state is temporally stable across sampling windows.
Persistence probability from stochastic models, which captures extinction risk.

The ICC approach is used in behavioral ecology and clinical research alike. Emodi-Perlman and colleagues pooled data from three ecological momentary assessment studies with 327 participants and over 32,000 reports to determine how many monitoring days were needed for reliable estimates of awake bruxism behaviors. Four of five behaviors reached excellent temporal stability after three monitoring days, with ICC values at or above 0.86, while grinding reached stability after two days with an ICC of 0.81 [6]. The lesson generalizes: stability estimates depend on sampling duration, and too short a window produces unreliable numbers.

Stability in Drug Storage and Pharmaceutical Science

In pharmaceutical science, stability is the ability of a drug product to maintain its identity, strength, quality, and purity within specified limits over time under defined storage conditions.

Chemical Stability and Shelf Life

Chemical stability means the active ingredient does not degrade into other compounds beyond acceptable limits. Shelf life is the period during which the product is expected to remain within specification. Degradation rate is the speed at which the active ingredient concentration falls, usually expressed as a percentage loss per unit time or as a rate constant.

The standard measure of drug stability is potency remaining over time, expressed as a percentage of the initial or labeled concentration. A common acceptance criterion is that the concentration must stay within a defined window, often plus or minus 10 percent of the label claim, for the product to be considered stable at a given time point.

ICH Storage Conditions

The International Council for Harmonisation (ICH) sets the standard storage conditions used in formal stability studies. Two are cited most often:

Long-term: 25 degrees C plus or minus 2 degrees C and 60 percent relative humidity (RH) plus or minus 5 percent RH.
Accelerated: 40 degrees C plus or minus 2 degrees C and 75 percent RH plus or minus 5 percent RH.

Long-term studies run for the proposed shelf life, often 12 to 24 months or longer. Accelerated studies typically run for six months and are used to predict degradation behavior and to justify short-term excursions outside the label conditions.

Soliman and colleagues applied exactly this framework to co-spray-dried theophylline dry powders for inhalation. Under accelerated conditions at 40 degrees C and 75 percent RH for three months, both formulations showed pronounced recrystallization with crystallinity up to 89.9 percent, agglomeration with particle sizes up to 217 micrometers, and deterioration in aerodynamic performance with fine particle fraction as low as 11.55 percent [7]. Under long-term desiccator storage at 25 degrees C for one year, the same formulations showed substantial recrystallization up to 80.7 percent while preserving thermal, chemical, and aerodynamic performance [7]. The same product can be stable by one measure and unstable by another, which is why stability protocols test multiple attributes.

Stability-Indicating Methods and Forced Degradation

A stability-indicating assay is an analytical method that can separate the active ingredient from its degradation products, so that a change in the measured peak reflects real degradation rather than assay noise. Daroi and colleagues developed and validated a stability-indicating reverse-phase HPLC method for mavacamten in oral capsules, achieving linearity from 5 to 75 micrograms per milliliter with a correlation coefficient of 0.9994 and recovery of 100.5 percent [8]. In forced degradation studies, they observed the highest degradation under acidic conditions at 10.37 percent, with minimal degradation under photolytic conditions at 0.93 percent, thermal at 1.66 percent, oxidative at 2.56 percent, and alkaline at 4.19 percent [8]. Forced degradation deliberately stresses the molecule to prove the method can detect degradation products that might form in real storage.

Product-Specific Stability

Stability depends heavily on the matrix. Majumdar and colleagues evaluated cannabinoid stability in cannabis plant material, extracts, oil formulations, and pure isolates of CBD and delta-9-THC under room temperature, refrigeration, and freezing over an extended period. Stability profiles varied depending on the cannabinoid profile, the temperature, and the matrix type [9]. The same molecule can be stable in one formulation and unstable in another.

Formulation can also protect a molecule. Wan and colleagues found that after 60 days of storage at 4, 25, 40, and 60 degrees C, the peroxide value of unencapsulated alpha-linolenic acid-rich structured lipid was 1.02, 1.06, 1.01, and 1.11 times higher than the microencapsulated version, showing that encapsulation improved oxidative stability [10]. In vaccine development, Gomes and colleagues found that squalene raw material degraded under all tested stress conditions, while formulation in nano-sized globules improved its stability, with degradation observed only under hydrogen peroxide and light exposure [11].

Storage temperature effects can be dramatic. Shields and colleagues tracked aggregation of the glycoprotein E antigen from a vaccine using stochastic electrochemical blocking and found that refrigeration at 4 degrees C consistently reduced aggregate size, while 37 degrees C accelerated aggregation until particles became too large to reach the electrode [12].

Biological matrix stability matters in forensic and clinical toxicology too. Riley-Carrier and Wagner assessed xylazine stability in preserved antemortem and postmortem blood at three storage conditions (23 degrees C, 2 to 8 degrees C, and minus 20 degrees C) at two concentrations, 3 nanograms per milliliter and 160 nanograms per milliliter. A sample was considered stable if the calculated concentration stayed within plus or minus 20 percent of the Day 0 value [13]. That 20 percent criterion is a common threshold in bioanalytical stability testing.

Summary Table: Stability Across Three Domains

Stability type	Definition	Key metric or unit	Example
Protein thermodynamic stability	Free energy difference between folded and unfolded states	Delta-G of unfolding in kcal/mol	Globular protein with ΔG of 5 to 15 kcal/mol at 25 degrees C [1]
Protein thermal stability	Temperature at which half the protein is unfolded	Tm in degrees Celsius	NanoLuc variants with increased thermostability from computational design [2]
Protein biological stability	Persistence of a protein inside a living cell	Half-life or degradation rate	HLTF rapidly degraded during HCMV infection via UL145 [3]
Ecological resistance	Magnitude of change caused by a disturbance	Percent change in abundance or composition	Vaginal microbiome composition shifts under ecological feedback [4]
Ecological resilience	Speed of return to pre-disturbance state	Return time in days or generations	Gut microecology returning to dynamic balance [5]
Ecological temporal stability	Consistency of a measurement across sampling windows	Intraclass correlation coefficient (ICC)	Awake bruxism behaviors stable after three monitoring days, ICC at or above 0.86 [6]
Drug chemical stability	Maintenance of active ingredient concentration within specification	Potency remaining as percent of label claim	Mavacamten capsules assayed by stability-indicating HPLC [8]
Drug physical stability	Maintenance of physical form and performance	Crystallinity percent, particle size in micrometers	Theophylline powders recrystallized up to 89.9 percent under accelerated conditions [7]
Formulation stability	Protection of active ingredient by the delivery matrix	Degradation rate or peroxide value	Microencapsulated ALA-SL had lower peroxide values than unencapsulated lipid [10]

How Stability Is Tested in Practice

Each domain uses a different experimental logic.

Protein chemistry uses denaturant titrations, thermal melts, and calorimetry. The output is a curve, and the parameters (ΔG, Tm, cooperativity) are extracted by fitting a model to that curve.

Ecology uses long-term monitoring, time series analysis, and stochastic modeling. The output is a probability or a rate, such as the probability that a population persists for another decade or the rate at which a community returns to baseline.

Pharmaceutical science uses formal stability protocols defined by ICH, with fixed time points, fixed storage conditions, and validated analytical methods. The output is a table of potency, impurity, and physical attribute values at each time point, from which shelf life is assigned.

Common Mistakes and Limitations

Treating stability as a single number. A protein with a high Tm can still be rapidly degraded in cells. A drug with excellent chemical stability can fail on physical attributes. Always ask which stability is being claimed.

Confusing thermodynamic stability with biological half-life. ΔG of unfolding describes folding equilibrium in a test tube. It says nothing about whether a cell will destroy the protein. The HCMV work on HLTF degradation shows how fast a functionally important protein can be removed despite being properly folded [3].

Assuming ΔG can be measured directly. The value at zero denaturant is an extrapolation from denaturant-dependent measurements, not a direct reading [1]. Reporting it without noting the model and the extrapolation is incomplete.

Ignoring the matrix in drug stability. The same molecule behaves differently in plant material, extract, oil, and isolate [9]. Formulation choices such as microencapsulation or nanoemulsion can change degradation rates substantially [10][11].

Using too short a sampling window in ecological or behavioral studies. Reliability depends on duration. Three days was enough for most awake bruxism behaviors but not all [6]. Extrapolating from a short window to a long-term claim is a common error.

Forgetting that storage temperature dominates. A shift from 4 degrees C to 37 degrees C can turn a stable formulation into an aggregating one [12]. Refrigeration is not a formality.

Assuming accelerated stability predicts long-term stability perfectly. Accelerated conditions at 40 degrees C and 75 percent RH can produce recrystallization and performance loss that do not appear under long-term storage at 25 degrees C [7]. Accelerated data guide decisions but do not replace long-term data.

Overlooking biological variability in stability studies. Stochastic models of microbial communities must account for random forces, ecological feedback, and sampling error separately, because each contributes differently to observed variability [4].

Individual cases, whether a specific protein construct, a specific ecosystem, or a specific compounded preparation, require expert judgment. A veterinarian, pharmacist, or principal investigator should be consulted for decisions that affect patient care or research conclusions.

Quick Review

Stability is context-dependent. It is always defined relative to a reference state, a set of conditions, and a time window.
Protein thermodynamic stability is measured by delta-G of unfolding in kcal/mol, typically 5 to 15 kcal/mol for globular proteins, and cannot be measured directly at zero denaturant [1].
Protein thermal stability is reported as Tm in degrees Celsius, the temperature at which half the protein is unfolded.
Ecological stability splits into resistance (how much change) and resilience (how fast recovery), plus persistence probability in stochastic models [4].
Drug stability is measured as potency remaining over time under ICH conditions, with long-term storage at 25 degrees C and 60 percent RH and accelerated storage at 40 degrees C and 75 percent RH [7].
Formulation and matrix change stability. Microencapsulation reduced peroxide formation in a structured lipid [10], and nanoemulsion improved squalene stability [11].
Sampling duration and storage temperature are two of the most common sources of error in stability measurements [6][12].

Frequently Asked Questions

What is the simplest definition of stability in science?

Stability is the tendency of a system to stay in or return to a defined state when conditions change. The defined state, the disturbance, and the observation window must all be specified for the definition to be useful.

Why is protein stability measured in kcal/mol?

Because protein stability is a free energy difference between the folded and unfolded states. Free energy is expressed in energy per mole, and kilocalories per mole is the conventional unit in biochemistry.

What does Tm mean for a protein?

Tm is the melting temperature, the point at which half the protein molecules are folded and half are unfolded. A higher Tm indicates greater tolerance for heat.

What is the difference between resistance and resilience in ecology?

Resistance is how little a system changes when disturbed. Resilience is how quickly it returns to its prior state. A system can be high in one and low in the other.

What are ICH storage conditions for drug stability testing?

The two most cited are long-term storage at 25 degrees C and 60 percent relative humidity and accelerated storage at 40 degrees C and 75 percent relative humidity. Long-term studies support shelf life, and accelerated studies predict degradation behavior.

Does a stable protein stay stable inside a cell?

Not necessarily. Cells actively degrade proteins through the proteasome and lysosome, so thermodynamic stability in a test tube does not guarantee persistence in a living system.