Genomic Informational Field Theory (GIFT) to characterize genotypes involved in large phenotypic fluctuations

Cyril Rauch,Panagiota Kyratzi,Andras Paldi

2023-07-05

Abstract:Based on the normal distribution and its properties, i.e., average and variance, Fisher works have provided a conceptual framework to identify genotype-phenotype associations. While Fisher intuition has proved fruitful over the past century, the current demands for higher mapping precisions have led to the formulation of a new genotype-phenotype association method a.k.a. GIFT (Genomic Informational Field Theory). Not only is the method more powerful in extracting information from genotype and phenotype datasets, GIFT can also deal with any phenotype distribution density function. Here we apply GIFT to a hypothetical Cauchy-distributed phenotype. As opposed to the normal distribution that restricts fluctuations to a finite variance defined by the bulk of the distribution, Cauchy distribution embraces large phenotypic fluctuations and as a result, averages and variances from Cauchy-distributed phenotypes cannot be defined mathematically. While classic genotype-phenotype association methods (GWAS) are unable to function without proper average and variance, it is demonstrated here that GIFT can associate genotype to phenotype in this case. As phenotypic plasticity, i.e., phenotypic fluctuation, is central to surviving sudden environmental changes, by applying GIFT the unique characteristic of the genotype permitting evolution of biallelic organisms to take place is determined in this case.

Populations and Evolution,Biological Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to process data sets with large phenotypic fluctuations through genotype - phenotype association analysis without relying on mean and variance?** Specifically, traditional genome - wide association study (GWAS) methods based on Fisher theory have limitations when dealing with complex traits, especially in cases of genetic variation with limited sample size or small effects. These problems lead to the "missing heritability" phenomenon, that is, the genetic variation identified through GWAS can only explain part of the phenotypic variation.

To solve these problems, the author proposes a new genotype - phenotype association method - **Genomic Informational Field Theory (GIFT)**. The core idea of GIFT is to use information - theoretic concepts such as information entropy, rather than the mean and variance in traditional statistical methods, to more accurately extract information from genotype and phenotype data sets. In addition, GIFT can handle any type of phenotypic distribution density function, not just the normal distribution.

### Specific problem description

1. **Limitations of traditional GWAS methods**:
- It depends on the normal distribution and its mean and variance.
- When dealing with complex traits (such as human height), due to the existence of small - effect genetic variation, a very large sample size is required to detect significant associations.
- For some phenotypic distributions (such as Cauchy distribution), traditional methods cannot define the mean and variance, so they cannot be effectively applied.

2. **The necessity of introducing GIFT**:
- GIFT can handle any phenotypic distribution, including those without a clear mean and variance (such as Cauchy distribution).
- GIFT provides more powerful tools to extract association information between genotypes and phenotypes through information - theoretic concepts such as information entropy.
- GIFT can provide more accurate genotype - phenotype mapping with a smaller sample size.

### Main contributions of the paper

1. **Propose a new genotype - phenotype association method (GIFT)**, which does not rely on mean and variance but is based on information - theoretic concepts.
2. **Demonstrate the effectiveness of GIFT in processing data sets with large phenotypic fluctuations**, especially for phenotypes that do not conform to the normal distribution.
3. **Explore the potential applications of GIFT in evolutionary biology**, for example, in the case of environmental mutations, promoting the evolution of survival characteristics through large phenotypic fluctuations.

Through these innovations, the paper aims to provide a more general and powerful framework for genotype - phenotype association analysis, so as to better understand the impact of genetic variation on phenotypes.