ADMET Analysis — The Test Every Drug Must Pass Before It Reaches You

In my last article, I wrote about molecular docking — how scientists test whether a drug molecule fits a target protein, like a key fitting a lock.

But here is something I did not tell you.

A perfect docking score does not mean the drug will actually work in a human body. A molecule can bind beautifully to a protein in a computer simulation and still be completely useless as a medicine — or worse, dangerous.

That is where ADMET analysis comes in.

What Does ADMET Even Stand For?

ADMET stands for:

A — Absorption D — Distribution M — Metabolism E — Excretion T — Toxicity

These five properties determine whether a drug can actually survive the journey through your body and do its job safely. A molecule that fails any one of these is eliminated from consideration — no matter how good its docking score was.

Think of ADMET as the real world test after the computer simulation. Docking tells you if the key fits the lock. ADMET tells you if the key can actually make it to the door without breaking, getting lost, or poisoning someone along the way.

Breaking It Down Simply

Absorption — Can the drug enter your bloodstream after you swallow it? A drug that cannot be absorbed orally is much harder to use practically. Scientists check something called GI absorption — whether the molecule can pass through your gut wall into your blood.

Distribution — Once in your blood, where does the drug go? Does it reach the right organs? One important check here is whether the drug crosses the blood-brain barrier. For brain diseases this is essential. For other conditions it can be a problem.

Metabolism — How does your liver process this drug? Your liver breaks down almost everything that enters your body. If it breaks down the drug too fast, it will not work long enough. If it cannot break it down at all, it may accumulate and become toxic.

Excretion — How does the drug leave your body? Through urine? Bile? If a drug cannot be excreted properly it builds up — and that is dangerous.

Toxicity — Is the drug safe? This includes checking for liver damage potential, mutagenicity, carcinogenicity, and how much of it would be lethal in animal models — expressed as LD50 values.

Lipinski’s Rule of Five — The Golden Standard

One of the most important filters in ADMET analysis is something called Lipinski’s Rule of Five. It was proposed by scientist Christopher Lipinski in 1997 and it remains one of the most widely used drug-likeness criteria in pharmaceutical research.

A drug-like molecule should have:

• Molecular weight of 500 g/mol or less
• 5 or fewer hydrogen bond donors
• 10 or fewer hydrogen bond acceptors
• LogP value of 5 or less

Molecules that violate more than one of these rules are unlikely to be orally bioavailable — meaning you cannot take them as a tablet or capsule. They would need injections or other delivery methods, which makes drug development far more complicated and expensive.

What I Found in My Own Research

For my final year research project at the University of the Punjab, I ran ADMET analysis on Cianidanol — the green tea flavonoid that showed the strongest binding affinity against alpha-amylase in my molecular docking study.

ADMET radar profile of Cianidanol showing favorable absorption, drug-likeness and low predicted toxicity. Generated using SwissADME (author’s own research, 2025).

I also analysed four structurally similar flavonoids for comparison.

Here is what the data showed for Cianidanol specifically:

ADMET profile of Cianidanol showing zero Lipinski violations and favorable bioavailability. Author’s own research, University of the Punjab (2025).

Zero Lipinski violations. Good bioavailability. Relatively easy to synthesize. These are genuinely encouraging results for a natural compound.

The one flag worth mentioning — the PAINS alert. PAINS stands for Pan-Assay Interference Compounds. A PAINS alert does not mean the compound is bad — it means it has structural features that can sometimes give false positives in lab assays. It is a signal for caution during future experimental testing, not a disqualification.

How Does Cianidanol Compare to Similar Flavonoids?

I also screened four structurally similar flavonoids. Here is a simplified comparison:

Comparison of molecular weight and synthetic accessibility scores across Cianidanol and four structurally similar green tea flavonoids. Lower synthetic accessibility score means easier and cheaper to produce as medicine. Data from author’s own research, University of the Punjab (2025).

All five compounds showed zero Lipinski violations and similar bioavailability scores — which confirms that this family of green tea flavonoids as a group has genuine drug-like potential.

Cianidanol stood out with the lowest molecular weight and the best synthetic accessibility score — meaning it would be relatively easier and cheaper to produce as a medicine compared to the others.

Why Does All of This Matter?

Here is the honest answer.

Most natural compounds with exciting biological activity never become medicines — not because they do not work, but because they fail ADMET. They cannot be absorbed. They are toxic. They are impossible to manufacture at scale.

Computational ADMET analysis — using free tools like SwissADME and ADMETlab — lets researchers screen thousands of compounds for these properties before spending years and millions on lab testing. It is not perfect. Computational predictions are not the same as clinical data. But they narrow the field dramatically and point research in the right direction.

When Cianidanol passed all of these filters in my study, it became not just an interesting docking result — but a genuinely promising drug candidate worth investigating further.

What This Means to Me

I ran this entire analysis on a computer. Free software. Publicly available protein databases. No expensive equipment.

And I found a green tea compound that outperformed FDA-approved drugs in computational testing and passed every drug-likeness filter.

That does not mean Cianidanol will become a diabetes drug. There is a long road from computational results to clinical trials. But it means the question is worth asking — and that is what science is.

Next week I will be writing about how natural compounds are being studied for cancer research — including some of the work from the Springer paper I co-authored earlier this year.

Follow me if you want to keep reading. 🍃

— Mamoona Nisar BS Chemistry (Biochemistry Major) | University of the Punjab Co-Author — In Silico Pharmacology, Springer (2026)