The discovery of the genome was one of the most important technological innovations of the last century. Fifty years after  Watson and Crick's description of DNA's fundamental structure the human genome has been decoded. "Today we are learning the language in which God created life."  said President Bill Clinton, proclaiming a new canon, The Book of Life, written in a  four letter language. It is somewhat unfortunate to reduce God's language to four letters. After all, according to the Jewish tradition (The Book of Creation), He created the world with 22 Hebrew letters. 

Does DNA language really describe life?  Or was it created  by geneticists to assist them in their research.  Like mathematics in the exact sciences? Geneticists regard it as the language of life.

Two models of disease

The DNA language  portrays  a simplistic model of a complex mechanism, that may be likened to a typewriter. While letter printing is well understood, all other myriad processes supporting it remain obscure. Medicine ought to take the DNA language seriously, since it claims to describe our daily bread, human disease. According to genetics, the genome consists of a complete set of instructions for making an organism. Diseases result from random changes (mutations) in this instruction set. Each disease and its faulty genes. It seems as if disease may be precisely specified, like molecules in chemistry.

Are genetic diseases identical with those treated by physicians? "Yes!" is the usual answer. All known diseases belong to a subset of the genetic. Genetics  and (clinical) medicine operate in two distinct realms. The domains of genotype and phenotype, and apply different models for describing disease.  Geneticists claim that both models are identical (iso-morphic). Each disease phenotype, corresponds  with its genotype. And yet when viewed from the perspective of the phenotype, diseases portrayed with the DNA language seem as caricatures of real diseases. The two models handle variation in a different way.

Gregor Mendel

Variation interfered in Gregor Mendel's experiments on garden peas, back in 1865.   Mendel suspected that heredity depended on contributions from both parents and that specific characteristics from each parent were passed on. He tested  the following model. Given two alleles that determine whether a seed is smooth (S) or wrinkled (s). And given that 'S' is dominant and 's' is recessive.  When a parent homozygous for the smooth seed allele (SS) is crossed with a parent homozygous for the wrinkled seed allele (ss), the gametes formed at fertilization have the genotype 'Ss' and the seed is spherical (1).

When the F1 plants self-pollinate they produce two kinds of eggs, and male sex cells , 'S' and 's'. These combine randomly in four different ways to form F2 plants ({SS}, {Ss}, {Ss}, {ss}).  Since S is dominant, three of the four possible combinations produce "spherical seed" phenotype, and the fourth produces "wrinkled seed" phenotype, so that the observed ratio ought to be 3:1.

Despite meticulous work, the observed smooth/wrinkled ratio deviated somewhat from the predicted. Instead of the expected 3:1, Mendel got 3.01:1, or 2.96:1, which might indicate that in some cases the model fails to explain observations, and has to be re-examined. In some plants, heredity  of seed pattern might be determined by additional factors, which will be called "Context" of the experiment. Mendel, a trained physicist, attributed variation to chance. The mean 3:1 ratio was genuine. Variation was random, and independent of the model.

Machine statistics

A similar approach is taken by physicists when examining machine attributes, like the ratio between area  and perimeter in circular disks. According to the predicted model, the ratio is pr2/2pr = r/2. In reality, disk ratios may vary, and  variation is attributed to chance. Variation fluctuates randomly about the mean. As more and more disks are measured variation becomes smaller, and vanishes when sample size is infinite.

Normal distribution

This behavior is expressed by a powerful model of machine statistics, the bell shaped curve (normal distribution). When the observed variables are independent of each other, they are distributed normally. Yet hardly any biological variable is distributed that way. While the normal distribution is symmetric and extends from minus-infinity to plus-infinity, biological distributions are asymmetric (skewed), and extend from zero to infinity. Had  Mendel examined his data, he might have realized that they were skewed. In asymmetric distributions variation cannot be attributed to chance alone.

Log-normal distribution

Despite such profound  conceptual difficulties, machine statistics dominate Epidemiology, and Medicine, which was achieved with a simple trick, called "transformation". Given a variable X, they conjured a constant 'a' so that log(X-a) is normally distributed. This log-normal distribution extends from zero to infinity, and approximates many biological distributions.  

Even this approximation is somewhat crude, and so are other transformations of machine statistics. None matches all observed (biological) data. Discrepancies are ignored and variation is still attributed to chance. In reality these discrepancies are not random.

These arguments are illustrated  by a simple equation.  Given a set of observations. Their behavior is explained as follows: 

Explanation = Model + Context.
Context accounts for the deviation of observed data from the model. In machine statistics context is regarded as a random error that can be minimized by enlarging the sample size, and the equation reduces to Explanation = Model. In biological distribution context does not vanish when sample size is infinite.

In many physical phenomena, like weather, variation cannot be reduced at will, and may even increase without a bound.  Like the butterfly effect, that characterizes a new branch of physics, known as non-linear dynamics, or Chaos, where machine statistics also fail. 

Genetics equates context with random error and ignores it, so that Genetic-disease = Model. Medicine on the other hand, refrains from such a simplification, and
Phenotypic-disease = Model + Context.

Gene tachycardia

The difference between the two disease models will be illustrated by a fictitious example of an imaginary genetic disease, called gene-tachycardia (GT). A random mutation changes the code of the GT-gene. It is transcribed to a faulty enzyme that fails to restrain heart rate, which  varies between 140 - 200/min.

The gene was discovered in individuals with tachycardia in an  otherwise  healthy heart. Genetic analysis of families with tachycardia, revealed that GT was inherited in a recessive mode. Parents were carriers (GT,gt), and their children inherited the four gene combinations ({GT,GT}; {GT,gt} ; {GT,gt} ; {gt,gt}). Only individuals with {gt,gt} had tachycardia. The same inheritance pattern that was discovered by Mendel in his peas.

Some time later the GT gene was sequenced. Closer analysis of the GT gene revealed about 150 different mutations at different locations, each associated with a different pulse rate. Some mutations do not affect the heart rate at all, while other are associated with rising tachycardia. A similar mutation variation is observed in many genetic diseases, e.g., thalassemia minor. 


It turned out that some {gt,gt} individuals had normal heart rate. Geneticists faced Mendel's dilemma, how to deal with "outliers" (variation)? Do these individuals undermine the validity of the model, or might a new interpretation of context (variation) save the model?

The solution was simple. Obviously not all genes are equally expressed in the phenotype, and therefore some {gt,gt} "outliers" had a normal heart rate.  Geneticists conjured a new variable "Penetrance", that controlls gene expression. At low penetrance, individuals appear healthy. After all they carry a genetic disease!

Penetrance and free will

Not all clinicians were aware of the genetic nature of some tachycardias.  A patient with pronounced GT (180/min) was examined by a clinician and since he was in good health, his tachycardia was ascribed to nervousness, or to a "weak heart" that might be strengthened with exercise. Patient enrolled in a fitness club and some time later his heart rate dropped to 160/min. As  more and more GT patients enrolled in fitness clubs, GT penetrance of the population declined. 

This example highlights the problematic aspects of the genetic disease model which takes variation as an act of chance, while in reality it involves also free will. By enrolling in a fitness club the patient actively interfered with the context of the above equation. The two disease models are not equivalent. The phenotype is more than the genotype, since it accounts also for free will.

Machines statistics ignore free will
Machines lack free will, so that their variation may be attributed to chance, and analyzed with machine statistics. The bell shaped curve is adequate only for dealing with machines, and fails whenever (non-random) free will interferes with an observed variable. This is why diseases portrayed with the DNA language are at best caricatures of real diseases. Yet Medicine fails to see it that way. Overwhelmed with the amazing discoveries of molecular biology, it adopts the genetic disease model with all it inconsistencies.

While Genetics regards variation as a nuisance, Medicine ought to regard it as treatment opportunity. Low penetrance indicates that "gene malfunctions" can be corrected, and penetrance driven even lower. Unfortunately medicine tends to ignore such treatment opportunities.

Breast cancer gene (BRCA)

Discovery of breast cancer genes followed along the same path as the fictitious GT gene.  First came the families with high breast cancer rate, and then the genes were discovered.  The official statistics are frightening (2):

"BRCA1 and BRCA2 cancer-predisposing mutations are inherited in an autosomal dominant manner. Each child of an individual with a cancer-predisposing mutation in the BRCA1 or BRCA2 gene has a 50% risk of inheriting the cancer-predisposing mutation" (2).

"The risk of any women developing breast cancer is 12.5% (1 in 8) and for developing ovarian cancer significantly less. However, in the presence of a germ-line BRCA-1 mutation these risks skyrocket to 85% and 50% for breast and ovarian cancer, respectively".

Are we doomed? Not at all, since these estimates were derived from "normalized" breast cancer frequency distributions (machine statistics), which regard variation as random noise. These estimates are extremely crude. 

BRCA penetrance

Their crudeness is revealed in BRCA penetrance estimates : 

"It appears that mutations to this gene only play a role in the development of approximately one-half of familial breast cancer, which accounts for 2-19% of all breast cancer, and about 10% of the cases of early onset breast cancer. That the BRCA-1 gene has not been found to be mutated in sporadic breast cancers" (2).

In other words, in familial breast cancer penetrance is less than 50%. Above all, most sporadic breast cancer patients do not carry this mutation. In the majority of breast cancer patients the model fails to describe variation adequately.  And such models  serve for risk estimates. This dangerous misconception convinced some genetic counselors to suggest  bi-lateral mastectomy for breast cancer prevention in so called "high risk" patients. 

Breast cancer is not a genetic disease! Gene mutations may accompany many (phenotypic) diseases, however disease progression is shaped by epigenetic factors, that emerge as disease evolves. Above all,  breast cancer patients can improve their prospect (and turn down their penetrance). By adopting the genetic model, Medicine ignores epigenetic opportunities for helping the patient (3).

Gene replacement

The gene-disease model breeds unjustified treatments, like gene replacement. The young and healthy female with a BRCA mutation carries a genetic disease, and should be saved from her ill fortune. Why not replace her faulty BRCA gene with a new one? In previous times such ideas were first tested on mice. Today more and more research is done on patients, which is unfortunate since the necessary requirements for replacing genes have not yet been met:

1.    Gene controls one process.
2.    Disease results from the malfunction (mutation) of one gene.
3.    The mutated gene can be targeted and replaced.

Most genetic diseases are poly-genic disorders, managed by gene families, and  mono-genic diseases are exceptions. Even in mono-genic diseases, gene replacement is still inconceivable, since mutations may alter different domains in the gene. Thalassemia minor  is hit by about 150 different mutations. BRCA is affected by about 100 different mutations.

Gene targeting is extremely crude. A vector carrying a gene cannot be inserted in the proper place of the genome. Instead it is applied to the entire genome and the gene is randomly incorporated at different sites, inducing undesired mutations with dangerous consequences.

These difficulties are highlighted in two mouse models, trans-gene mouse, and knock-out mouse. A trans-gene mouse carries foreign DNA in its genome. Even if it is a single gene, the phenotype may vary and the outcome is generally unpredictable.

Knock-out mice were developed for investigating the outcome of a single gene destruction. Even the knocking-out of a single gene produces unpredictable variation in the phenotype. In both experimental models none of the above pre-requisites is met.

Genetics and Ethics

Recent discoveries in genetics have far reaching ethical legal, and social implications, like : "Should a person with a genetic predisposition to vertigo be permitted to become an airline pilot? Or should a middle-aged person with a genetic predisposition to colon cancer be denied employment because that person would likely need expensive medical care that would raise health insurance premiums for everyone at the company?"

Such issues seem insurmountable provided that one accepts the simplistic genetic model of disease, where a person with genetic vertigo carries a disease.  In reality he is not sick and may help himself. By banning free will from its models, genetics portrays a fatalistic universe of diseases. Genes determine ones destiny just like the stars in ancient Babylon. One of your ancestors was hit by a random mutation. You inherited this ill fate and transmit it to your progeny.

On the other hand, without the banning of free will from its models, genetics would never have made its amazing discoveries.

Many so called ethical issues arose since society decided to accept the genetic view of life, and ignores context. The "unfortunate" person with a homozygous gene for vertigo, may still improve his prospects by enrolling  an "anti-vertigo fitness club" and become an astronaut. Even then genetics would regard him as sick since carrying a faulty gene. Should we also regard him that way?

Bell shaped curve

A similar misconception bred a best seller called  The Bell Curve (4), according to which intelligence varies among individuals, and these differences profoundly influence the social structure in modern societies. The authors adopt the fatalistic view of genetics and base their conclusions on machine statistics. The cognitive measure IQ, is supposedly distributed as the  bell shaped curve. Intelligence is in one's genes. An individual that inherited low-intelligence genes cannot escape his misfortune, and will path it to his progeny.

The bell shaped distribution is a false model for expressing IQ variation. Since it extends from minus-infinity to plus-infinity, it predicts the existence of individuals with negative IQ, which in itself suffices for discrediting it. Genetics reveals that men may not be created equal, yet ignores the fact that their phenotype is creative, otherwise life would not be possible. 

And God created the world using 22 Hebrew letters. Four were reserved for the genetic code, and the rest, for free will.


1. Mendel G. Experiments in Plant Hybridization (1865)

2. Bars Culver J, Hull J, Levy-Lahad, E, Daly M,Burke W,
Breast Cancer Genetics - An Overview

3. Zajicek G. Cancer and Wisdom of the Body.

4. Herrnstein R.J, Murray C. The Bell Curve: Intelligence and Class Structure in American Life
New York: Free Press, 1994. 845 pp. ISBN 0-02-914673-9.

Click on your icon to follow your trail


Back to New Medicine