The accurate prediction of biological features from genomic data is paramount for precision medicine, sustainable agriculture and climate change research. For decades, neural network models have been widely popular in fields like computer vision, ast
rophysics and targeted marketing given their prediction accuracy and their robust performance under big data settings. Yet neural network models have not made a successful transition into the medical and biological world due to the ubiquitous characteristics of biological data such as modest sample sizes, sparsity, and extreme heterogeneity. Results: Here, we investigate the robustness, generalization potential and prediction accuracy of widely used convolutional neural network and natural language processing models with a variety of heterogeneous genomic datasets. While the perspective of a robust out-of-the-box neural network model is out of reach, we identify certain model characteristics that translate well across datasets and could serve as a baseline model for translational researchers. Here, we investigate the robustness, generalization potential and prediction accuracy of widely used convolutional neural network and natural language processing models with a variety of heterogeneous genomic datasets. While the perspective of a robust out-of-the-box neural network model is out of reach, we identify certain model characteristics that translate well across datasets and could serve as a baseline model for translational researchers.
WI Fast Stats is the first and only dedicated tool tailored to the WI Fast Plants educational objectives. WI Fast Stats is an integrated animated web page with a collection of R-developed web apps that provide Data Visualization and Data Analysis too
ls for WI Fast Plants data. WI Fast Stats is a user-friendly easy-to-use interface that will render Data Science accessible to K-16 teachers and students currently using WI Fast Plants lesson plans. Users do not need to have strong programming or mathematical background to use WI Fast Stats as the web apps are simple to use, well documented, and freely available.
The Tree of Life is the graphical structure that represents the evolutionary process from single-cell organisms at the origin of life to the vast biodiversity we see today. Reconstructing this tree from genomic sequences is challenging due to the var
iety of biological forces that shape the signal in the data, and many of those processes like incomplete lineage sorting and hybridization can produce confounding information. Here, we present the mathematical version of the identifiability proofs of phylogenetic networks under the pseudolikelihood model in SNaQ. We establish that the ability to detect different hybridization events depends on the number of nodes on the hybridization blob, with small blobs (corresponding to closely related species) being the hardest to be detected. Our work focuses on level-1 networks, but raises attention to the importance of identifiability studies on phylogenetic inference methods for broader classes of networks.
