Biophysicist Christoph Adami’s book The Evolution of Biological Information miscalculates viral evolution and information theory by tracking shifting treatment protocols rather than actual viral adaptation, according to an analysis by science critic and researcher Richard Lenski.
Flawed Premises in HIV Protease Inhibitor Analysis
Section 3.3 of Adami’s book, titled “Information Loss and Gain in HIV Evolution,” argues that introducing a protease inhibitor initially causes a decrease in information because viral genomes no longer match the altered environment. According to Adami’s theoretical model, as HIV adapts to the drug, it accumulates information about the presence of the inhibitor.
Adami supports this conclusion using patient-derived HIV genomes, noting an initial increase in sequence entropy during treatment, followed by a steady increase in information over time. However, Lenski points out a fundamental flaw in this population premise: patients undergoing treatment do not represent an isolated, independent population. Instead, they constitute a continuous stream of newly infected individuals predominantly carrying wild-type HIV. Consequently, the data measures changing clinical treatment protocols over the years rather than real-time viral evolution, according to Lenski’s critique.
Mathematical Inconsistencies and Sequence Entropy Approximations
To calculate sequence entropy, Adami initially assumes each position in a genetic sequence operates independently. When this simplified approach fails to show the expected increase in information over time—instead showing a decrease—Adami adopts a more sophisticated approximation that accounts for pairwise dependencies between sites.
Lenski identifies a major mathematical contradiction in this second approach. Analyzing 76 positions should yield a maximum theoretical capacity of 76 mers (where a mer represents maximum information for a single position). Yet, Adami’s calculations produce total information values fluctuating between 74 mers and 80 mers, placing most sequences right at or above their absolute theoretical limits.
Did you know?
When drug-resistant mutations emerge in viral genomes, they often compromise the virus’s structural stability, requiring compensatory mutations at distant sites to restore function—a phenomenon that simple pairwise calculations fail to capture.
Higher-Order Correlations and Sequence Variability
Adami acknowledges that his method omits higher-order correlations because data availability is too limited to estimate them reliably. He writes that researchers must simply “hope that this contribution is small,” a presumption Lenski disputes.
Resistance mutations in HIV are frequently damaging on their own, requiring numerous compensatory mutations that establish widespread correlations across multiple sites in the genome. Rather than tracking an increase in biological information, Adami’s altered methodology actually captures a rise in sequence variability. As modern treatment regimens expand, the virus explores a wider array of escape pathways, resulting in a more diverse viral population rather than a net gain in informational content.
Frequently Asked Questions
What does Christoph Adami’s book argue about HIV and information?
Adami argues that introduction of protease inhibitors initially destroys genomic information by mismatching the virus from its environment, while subsequent drug resistance represents an accumulation of information about the drug-laden environment.
Why does Richard Lenski argue Adami’s conclusions are incorrect?
According to Lenski, Adami’s data measures shifting clinical treatment protocols in a stream of newly infected patients rather than viral evolution within an independent population, and relies on flawed mathematical approximations that exceed theoretical information limits.
How do protease inhibitors affect HIV sequence variability?
Protease inhibitors drive an increase in sequence variability as the virus utilizes various mutational pathways to develop drug resistance, according to the critical analysis.
Join the Discussion: What are your thoughts on how theoretical models measure biological information? Share your perspective in the comments below, explore our archive for more evolutionary biology breakdowns, or subscribe to our newsletter for the latest updates.
Related reading