Rampant False Detection of Adaptive Phenotypic Optimization by ParTI-Based Pareto Front Inference

Abstract

Organisms face tradeoffs in performing multiple tasks. Identifying the optimal phenotypes maximizing the organismal fitness (or Pareto front) and inferring the relevant tasks allow testing phenotypic adaptations and help delineate evolutionary constraints, tradeoffs, and critical fitness components, so are of broad interest. It has been proposed that Pareto fronts can be identified from high-dimensional phenotypic data, including molecular phenotypes such as gene expression levels, by fitting polytopes (lines, triangles, tetrahedrons, and so on), and a program named ParTI was recently introduced for this purpose. ParTI has identified Pareto fronts and inferred phenotypes best for individual tasks (or archetypes) from numerous data sets such as the beak morphologies of Darwin’s finches and mRNA concentrations in human tumors, implying evolutionary optimizations of the involved traits. Nevertheless, the reliabilities of these findings are unknown. Using real and simulated data that lack evolutionary optimization, we here report extremely high false-positive rates of ParTI. The errors arise from phylogenetic relationships or population structures of the organisms analyzed and the flexibility of data analysis in ParTI that is equivalent to p-hacking. Because these problems are virtually universal, our findings cast doubt on almost all ParTI-based results and suggest that reliably identifying Pareto fronts and archetypes from high-dimensional phenotypic data are currently generally difficult.

Pareto optimality, economy–effectiveness trade-offs and ion channel degeneracy: improving population modelling for single neurons

“…Aside from that, ParTI is based on statistical inference of Pareto optimality [99] comparing the set of solutions to randomly shuffled data under the assumption that phenotypic traits are independent and the data are uncorrelated. It has been pointed out that this may not be the case in biological datasets due to phylogenetic correlations of traits [111,112] (see also [34]). This would make the identification of Pareto fronts in high-dimensional datasets prone to errors.…”

Neural Optimization: Understanding Trade-offs with Pareto Theory

“…Specifically, fitted polytopes to phenotypic data sets have been tested for significance based on a t-ratio test [63]; the fit to the polytope is compared with a shuffled version to estimate significance. Shuffling the data, however, assumes that traits are uncorrelated and the method, therefore, tends to overestimate significance [63,64]. objectives forms a straight line between the two archetypes (numbered dots) under certain assumptions [1].…”

Propagating uncertainty about molecular evolution models and prior distributions to phylogenetic trees

“…The ability of researchers to select arbitrary data analysis methods in order to make results appear more certain is already a recognized problem in evolutionary biology (Nakagawa et al, 2021;Sun and Zhang, 2021). The problem occurs in molecular phylogenetics whenever biologists select Bayesian methods because they make phylogenies appear more certain; that practice is widespread according to Bromham (2016, p. 431).…”