Reviews & Analysis

The field of biomolecular modeling has thrived by exploiting state-of-the-art technological advances. In this Perspective, the role of software and hardware advances, and the disparity and synergy between knowledge-based and physics-based methods are discussed and explored.

Mapping X-ray diffraction patterns to crystal structures is a comprehensive and time-consuming task for chemists and materials scientists. In a recent work, researchers developed a machine-learning tool to make this job more ‘self-driving’.

In this issue, a protocol for querying criminal DNA databases is developed to prevent profile discovery, therefore reducing the potential for racial bias.

Analyzing cellular barcoding data is challenging due to the absence of a flexible, complete and easy-to-use set of tools that can help scientists derive biological meaning from these datasets. barcodetrackR is a promising package for filling this gap in the field.

A limitation of data obtained from RNA-seq experiments is the presence of different types of cell expression, making it difficult to identify the contribution of cell-type composition or cell-type-specific expression. A new study addresses this problem by proposing a method for cell-type-aware analysis of RNA-seq data.

The computational complexity of deep neural networks is a major obstacle of many application scenarios driven by low-power devices, including federated learning. A recent finding shows that random sketches can substantially reduce the model complexity without affecting prediction accuracy.

Sharing of genomic data poses a problem due to privacy concerns and lack of individual’s choice in the matter. In this issue, researchers propose a framework for sharing data where the control lies with the individual.

A model for the electrical double layer at solid-state electrochemical interfaces is reported, shedding some light on the design and optimization of future all-solid-state Li-ion batteries.

The advent of STED microscopy, which allows observation at a sub-diffraction resolution, raises a challenge in studying spatial proximities of biomolecules’ distributions. In this issue, researchers have attempted to study colocalization of molecules by employing optimal transport.

Massive datasets have been made available to enable systematic studies of gene regulation and its control via epigenetic mechanisms. In this Review, state-of-the-art computational methods used to effectively extract knowledge from these datasets are presented and discussed.

The identification of genes that are associated with developmental trajectories of cells is an important focus in single-cell transcriptomics. This issue presents Scellnetor, a resource that facilitates this task.

Understanding and quantifying the uncertainty of predictions from COVID-19 pandemic models is essential to inform public health decision making. This issue presents one such examination using the influential CovidSim model.

Obtaining a consistent taxonomy of neuron types is challenging mainly because of the high dimensionality of the datasets. Coupled autoencoders are a step forward in achieving this goal.

Quantum computing has the potential to assist with myriad tasks in science. In this Perspective, the applicability and promising directions of quantum computing in computational biology, genetics and bioinformatics is evaluated and discussed.

There have been substantial developments in weather and climate prediction over the past few decades, attributable to advances in computational science. The rise of new technologies poses challenges to these developments, but also brings opportunities for new progress in the field.

The mechanisms facilitating evolutionary adaptation to future challenges are difficult to establish experimentally. Recent computational simulations of 200 cell populations indicate how evolution can hide useless genetic switches with capacity for later use.

The nature of biological networks still brings challenges related to computational complexity, interpretable results and statistical significance. Recent work proposes a new method that paves the way for addressing these issues when analyzing cancer genomic data.

Characterizing the aggregation of the peptide amyloid β is essential to better understand Alzheimer’s disease and to find potential targets for drug development. Deep neural networks make it possible to describe the kinetics of this peptide, opening the way for achieving this goal.

Computational approaches for drug repurposing can accelerate the identification of treatments during a pandemic. In this Review, the authors discuss this topic in the context of COVID-19 and propose a strategy to make computational drug repurposing more effective in future pandemics.

While estimating causality from observational data is challenging, quasi-experiments provide causal inference methods with plausible assumptions that can be practical to a range of real-world problems.