Search

Cell biology is fundamentally limited in its ability to collect complete data on cellular phenotypes and the wide range of responses to perturbation. Areas such as computer vision and speech recognition have addressed this problem of characterizing unseen or unlabeled conditions with the combined advances of big data, deep learning, and computing resources in the past 5 years. Similarly, recent advances in machine learning approaches enabled by single-cell data start to address prediction tasks in perturbation response modeling. We first define objectives in learning perturbation response in single-cell omics; survey existing approaches, resources, and datasets (https://github.com/theislab/sc-pert); and discuss how a perturbation atlas can enable deep learning models to construct an informative perturbation latent space. We then examine future avenues toward more powerful and explainable modeling using deep neural networks, which enable the integration of disparate information sources and an understanding of heterogeneous, complex, and unseen systems.

Single-cell transcriptomics is a versatile tool for exploring heterogeneous cell populations, but as with all genomics experiments, batch effects can hamper data integration and interpretation. The success of batch-effect correction is often evaluated by visual inspection of low-dimensional embeddings, which are inherently imprecise. Here we present a user-friendly, robust and sensitive k-nearest-neighbor batch-effect test (kBET; https://github.com/theislab/kBET) for quantification of batch effects. We used kBET to assess commonly used batch-regression and normalization approaches, and to quantify the extent to which they remove batch effects while preserving biological variability. We also demonstrate the application of kBET to data from peripheral blood mononuclear cells (PBMCs) from healthy donors to distinguish cell-type-specific inter-individual variability from changes in relative proportions of cell populations. This has important implications for future data-integration efforts, central to projects such as the Human Cell Atlas.

Scanpyis a scalable toolkit for analyzing single-cell gene expression data. It includes methods for preprocessing, visualization, clustering, pseudotime and trajectory inference, differential expression testing, and simulation of gene regulatory networks. Its Python-based implementation efficiently deals with data sets of more than one million cells (https://github.com/theislab/Scanpy). Along with Scanpy, we present AnnData, a generic class for handling annotated data matrices (https://github.com/theislab/anndata).

The temporal order of differentiating cells is intrinsically encoded in their single-cell expression profiles. We describe an efficient way to robustly estimate this order according to diffusion pseudotime (DPT), which measures transitions between cells using diffusion-like random walks. Our DPT software implementations make it possible to reconstruct the developmental progression of cells and identify transient or metastable states, branching decisions and differentiation endpoints.

Ion implantation of boron is a promising technique for the preparation of p-type emitters in n-type cells. We use rapid thermal annealing with temperatures up to 1250 °C and annealing durations between 6 s and 20 min to anneal the implant-induced crystal defects. Experimental J0e is compared with simulated and measured defect densities. Perfect dislocation loops are identified to be the dominating defect species after rapid thermal annealing (RTA) above 1000 °C. Even for emitters with J0e values around 40 fA/cm2, defects are present within the valleys of the textured surfaces after annealing. On textured Al2O3 -passivated boron emitters, we measure J0e of 38 fA/cm 2 for a sheet resistance around 80 Ω/□ after very short annealing processes (1 min at 1200 °C).

We present a fully calibrated model for the diffusion, segregation, and activation of phosphorus for typical annealing conditions of implanted silicon solar cells. In contrast to existing process simulation software, this model allows the quantitative prediction of doping profile distributions, and, thereby, sheet resistances, surface concentrations, and junction depths. The model also provides an intuitive understanding of the dependence of these quantities on the parameters of the annealing process.

This paper is motivated by the question of how residual implantation damage degrades solar cell performance. In order to avoid such degradation, annealing processes of implanted c-Si solar cells use high thermal budgets. Still, implantation-induced dislocation loops may survive these processes. We derive two models for the annealing kinetics of dislocation loops that are suitable for the study of high thermal budgets: a model that is able to describe the parallel ripening of faulted and perfect dislocation loops and a model that explicitly implements the conservative and nonconservative processes associated with Ostwald ripening. Both models lead to a better agreement with the experiment than what has been published before.