About this Research Topic
In recent years, new knowledge on molecular genetics and the rapid evolution of sequencing and genotyping technology has renewed the interest on genetic prediction. It should be recalled, however, that Mendel gave the basic model to predict traits determined by a single gene in the 19th century. The model was extended to complex traits, affected by several genetic and environmental factors, in the beginning of the 20th century by, among other, Fisher, Haldane and Wright. Since then, the prediction of complex traits has been a traditional field in epidemiology and also in animal and plant breeding.
Nowadays, genetics and genomics continue to evolve rapidly. The possibility of observing genomes, exomes and transcriptomes was rapidly followed by the on-going description of families of non-coding DNA. At the same time, spatial and temporal dynamics of the genome is being characterized by studies on transposable elements. All these DNA descriptors, based on nucleotide sequences, are being completed with new information on other transmissible non-nucleotide characteristics of chromatin.
Many biology branches are benefitting from this knew knowledge. Contributions are expected in individualized medecine, specialized medical applications such as HLA immputation and supporting decision making in transplantations, pharmaco and nutrigegenomics, and bio-industrial and agricultural applications. In this Research Topic we concentrate on the individual variability of these genetic descriptors, i.e. DNA Polymorphism, and its value to predict phenotypes. In particular we gather original contributions and review articles based on various "genetic models", using different kinds of "data", developing new "statistical approaches" and studying "reporting and ethical aspects", as described below.
All genetic models are considered in this Research Topic, from simple mendelian to complex traits. It is assumed that the exploitation of acquired knowledge, such as the identification of causal mutation or regions harboring QTL, improves genetic prediction. Models considering gene by gene interactions are relevant for the prediction of phenotypes. Both knowledge-based models and black-box models are included. We focus solely on the impact of individual variability on phenotypes because environmental effects are already discussed elsewhere.
In relation to Data, many authors have reported the contribution of DNA marker data (microsatellites and then SNP) and RNA expression data to genetic prediction, but the complementary value of polymorphism of CNV, non-coding-DNA and chromatin characteristics has been seldom reported.
We also review statistical methodology by gathering many statistical approaches being proposed for genetic prediction including Maximum likelihood, Bayesian inference and Machine Learning tools. In addition, meta-methods which optimally combine several predictors are emerging as a useful tool to combine data of diverse nature in the process of prediction.
Reporting and ethical aspects. It is important that the predicted result be coupled to some measure of the confidence given to each individual prediction. This is particularly true for complex traits where it is impossible to provide a deterministic prediction. This aspect, traditionally well covered in human genetics and also in plant and animal breeding, is sometimes being neglected in some commercial services for genetic prediction. Here, we encourage authors to contribute on ethical issues about data handling and the reporting of genetic predictions.
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.