Learning by Subtraction: How Human Biology Reveals What We Can Live Without
Medicine cannot remove human functions simply to discover which ones matter. Nature, disease and accident nevertheless produce cases that resemble experiments in subtraction.
Some people carry damaging variants in both copies of a gene. Others develop without an organ, lose one through disease or surgery, or survive injuries that remove substantial parts of the body. Modern medicine creates another category by transferring biological work to drugs, prostheses and machines.
These cases offer an empirical route into the question of what a human system actually requires. But survival is an ambiguous result. A person who lives without a structure has not necessarily shown that it was useless. Another pathway may have absorbed its work, development may have reorganised around its absence, or medicine may be supplying enough of the missing function to keep the larger system operating.
The cost may also remain hidden until infection, injury, pregnancy, ageing or some other form of stress exposes a loss of reserve. What survives subtraction is not an ordinary human with one component neatly removed. It is an organism that has compensated, adapted or become dependent on support.
The useful question is therefore not which human parts are unnecessary. It is which functions can disappear, which can be redistributed and which must survive in another form.
Natural Loss of Function
Laboratory biologists often investigate a gene by disabling it. If a cell or animal changes in a consistent way, the missing gene’s contribution becomes easier to identify. Humans cannot ethically be studied by deliberately deleting genes, but naturally occurring genetic variation sometimes produces a related form of evidence.
Some people inherit predicted loss-of-function variants in both copies of a gene. These individuals have often been described as human knockouts, although the term is less exact than it sounds. A sequence variant predicted to disrupt a protein may leave some activity intact, affect only certain transcripts or be incorrectly annotated. Establishing genuine loss of function requires more than finding a premature stop signal in a database.
A 2017 study of 10,503 adults in Pakistan identified homozygous predicted loss-of-function variants affecting 1,317 genes in at least one participant. The population’s high degree of familial relatedness increased the likelihood that rare variants inherited from a common ancestor would occur in both copies of a gene. Researchers then compared those genotypes with more than 200 biochemical and disease-related traits.
The project has since expanded dramatically. The Pakistan Genome Resource reported in June 2026 contains exome or genome data from 173,303 participants and identifies naturally occurring homozygous loss-of-function variants in 6,476 genes—roughly one-third of human protein-coding genes.
The newer study also adopts more cautious terminology. “Knockout” normally describes a purposeful experimental deletion, whereas these participants carry naturally occurring variants predicted to disable both copies of a gene. The distinction is more than politeness. It keeps the evidence aligned with what the sequencing has actually shown.
Even a convincingly inactivated gene does not produce a clean before-and-after experiment. The person developed from conception under those conditions. Other pathways may have adjusted during embryonic development, childhood or adulthood. The observed individual is not an ordinary person minus one protein, but an organism assembled under a different constraint.
That makes the adaptation as informative as the absence. Natural loss-of-function cases can reveal what a gene contributes, but they also reveal how much flexibility surrounds its loss.
When Absence Identifies a Drug Target
Some natural genetic differences have done more than clarify biological function. They have shown where a pathway might be altered deliberately.
The best-known example is PCSK9, a protein involved in regulating the receptors that remove LDL cholesterol from the blood. The original human genetic studies found that people carrying particular PCSK9 variants had lower lifelong LDL cholesterol and substantially lower rates of coronary heart disease.
Those observations helped validate PCSK9 as a therapeutic target. Injectable PCSK9 inhibitors later became established cholesterol-lowering treatments, and in July 2026 the United States Food and Drug Administration approved the first oral PCSK9 inhibitor.
This is an unusually successful example of learning from subtraction. The natural variants suggested that reducing the pathway could lower cardiovascular risk without creating an obvious catastrophic deficiency. Medicine could then attempt a controlled and partial version of the same change.
The analogy remains imperfect. Someone born with reduced PCSK9 activity has lived with it from the beginning. A drug introduced at forty or seventy acts on a body developed under ordinary PCSK9 signalling. Inherited absence and pharmaceutical inhibition may involve the same pathway without constituting the same biological intervention.
Timing matters in both directions. A protein may be essential while an embryo is forming but less important in adulthood. Alternatively, lifelong development may construct compensatory pathways that are unavailable when an adult loses the same function abruptly. A natural loss-of-function variant can therefore show that a human organism can sometimes be built under a particular constraint. It does not automatically show that an existing organism can safely acquire that constraint later.
Four Ways to Survive Without Something
The phrase “a person can live without it” hides several different biological situations. Treating them as interchangeable turns useful evidence into an inventory of supposedly expendable parts.
Genuine dispensability would be the strongest case: loss of a component causes no meaningful decline in health, capacity or resilience across ordinary life and physiological stress. This category may exist, but it is difficult to prove. A function that appears irrelevant under routine conditions may become important during an uncommon infection, injury, pregnancy or environmental exposure.
Redundancy means that another pathway performs overlapping work. Removing one component has little immediate effect because the surviving pathway absorbs the demand. Redundancy may look like biological inefficiency, but it often provides reserve. The missing component was not useless; it helped create the margin between ordinary operation and system failure.
Replacement allows anatomy to disappear while preserving enough of its function elsewhere. Dialysis does not demonstrate that kidney function is unnecessary. It demonstrates that selected tasks can be transferred to machinery well enough to sustain life. Hormone replacement can make an endocrine organ anatomically removable while confirming the importance of the signal it once produced.
Developmental bypass occurs when an organism grows around an absence. Neural circuits reorganise, another organ enlarges or regulatory systems adopt new operating points. The resulting person may function remarkably well, but only because development built an alternative route from the beginning.
These categories overlap. A redundant pathway may become stronger during development, while an externally replaced organ may have secondary functions that medicine reproduces poorly. The purpose of the distinction is not to classify every case perfectly. It is to recognise that continued life can be achieved through several different mechanisms.
Organ loss makes the problem concrete. Most people with one functioning kidney can live full and healthy lives, but a solitary kidney still warrants protection and monitoring because reduced function or high blood pressure can emerge later. The second kidney provided reserve even when the first appeared capable of handling ordinary demand alone.
A person can also live without a functioning spleen, but the loss changes their vulnerability. The CDC identifies anatomic or functional asplenia as a risk factor for serious infection by encapsulated bacteria and recommends additional vaccination. Survival does not mean the spleen performed no important work; it means that precautions and other immune defences can sometimes keep the increased risk manageable.
What organ loss proves is therefore conditional. It shows that the larger system can continue under particular circumstances. It does not show that the same system retains all of its former resilience.
The Brain and the Survivorship Filter
The brain provides the most dramatic examples and the greatest temptation to confuse adaptation with spare capacity.
Some people who underwent hemispherectomy during childhood—removal or functional disconnection of one cerebral hemisphere, usually to control severe epilepsy—later developed language, cognition and social abilities far better than the quantity of remaining tissue might suggest.
A 2019 study of six adults who had undergone the procedure during childhood found that familiar large-scale functional networks remained identifiable within the surviving hemisphere. Connectivity between some different networks was stronger than in control participants, suggesting extensive reorganisation within the remaining tissue.
The result does not demonstrate that half of a typical adult brain is unnecessary. The operations occurred during childhood, when neural systems retain substantial developmental plasticity. The participants were also unusually high-functioning survivors selected for an intensive imaging study, not a representative sample of every hemispherectomy outcome.
The evidence concerns plasticity rather than anatomical minimalism. A developing brain with far less than its expected structure may still construct many necessary networks if enough tissue, time and favourable conditions remain. The adult who emerges is not the original child with half a brain simply subtracted. Years of development have produced a different organisation.
The same selection problem affects genetic databases. Researchers can study only people who survived long enough to be sequenced. A homozygous loss-of-function variant found in a healthy adult is necessarily compatible with development and survival in at least one genetic and environmental setting. Variants that prevent embryonic development or cause early death will be systematically absent from adult cohorts.
The 2020 gnomAD constraint analysis, based on variation in 141,456 people, approaches this problem by comparing how many disruptive variants are observed in each gene with how many ordinary mutation rates would predict. Genes with far fewer loss-of-function variants than expected show evidence of constraint, consistent with selection removing damaging variants from the observable population.
Constraint is not a simple division between essential and non-essential genes. Genes occupy a continuous spectrum. Statistical power also varies with gene size, and even a database of more than 140,000 people remained unable to assess constraint confidently for a substantial minority of coding genes. Effects that emerge only after reproductive age may also leave a weaker evolutionary signature because they do less to prevent the variant from being passed on.
The map of what humans can tolerate must therefore combine two forms of evidence: components found missing in living people and components conspicuously absent from those people. The first suggests tolerance under observed conditions. The second suggests biological constraint, although sequencing limitations and inadequate sample sizes can also produce absence.
Phenotyping creates another limit. “No recognised effect” may mean that the missing gene causes no meaningful difference. It may also mean that researchers did not measure the uncommon infection, fertility outcome, cognitive trait, injury response or late-life condition capable of revealing one.
A catalogue of genetic absence is only as complete as the catalogue of human experience against which it is compared.
From an Inventory to a Dependency Map
A serious attempt to map biological necessity would need evidence at several scales. Genomic resources could identify people carrying disruptive variants in both copies of particular genes. Medical records and long-term follow-up could reveal disease, fertility, survival and late-life outcomes. Deep phenotyping could examine metabolism, immunity, cognition and responses to unusual forms of stress.
Congenital organ absence, surgical removal and traumatic injury would add anatomical evidence. Dialysis, artificial circulation, hormone replacement and prosthetic systems would show which functions can be transferred. Rehabilitation research would reveal how much apparent tolerance depends upon gradual adaptation rather than immediate replacement.
The result would not be a list entitled Parts a Human Does Not Need. It would be a dependency map. For each component, it would record when the function becomes necessary, which systems compensate for its loss, how much reserve disappears, which environments expose the deficit and whether medicine can replace the function without reproducing the original anatomy.
Such a map could identify promising drug targets, reveal overlooked biological roles and guide the design of artificial organs. It could distinguish structures that must be reproduced faithfully from functions that can be supplied more abstractly.
It would also make the minimal human harder to define. The smallest viable system cannot be described only by counting the biological parts that remain. It includes developmental changes, reserve pathways, medicines, machines, monitoring and the social infrastructure that keeps those supports available.
A person may survive without an organ while depending upon a medical system larger and more complicated than the organ that was lost. A gene may disappear only because another pathway has inherited its work. A childhood hemispherectomy may preserve function because years of development reorganised the remaining brain around the absence.
The human body is remarkably tolerant of subtraction, but tolerance is not emptiness. Survival after loss reveals the depth of the surrounding system and the work transferred to everything that remains.
The minimal human is not whatever is left after enough parts have been removed. It is whatever is left together with everything that has reorganised, compensated or been built around the loss.
Comments
Post a Comment