While only recently developed, the ability to profile expression data in single cells (scRNA-Seq) has already led to several important studies and findings. However, this technology has also raised several new computational challenges. These include questions about the best methods for clustering scRNA-Seq data, how to identify unique group of cells in such experiments, and how to determine the state or function of specific cells based on their expression profile. To address these issues we develop and test a method based on neural networks (NN) for the analysis and retrieval of single cell RNA-Seq data. We tested various NN architectures, some of which incorporate prior biological knowledge, and used these to obtain a reduced dimension representation of the single cell expression data. We show that the NN method improves upon prior methods in both, the ability to correctly group cells in experiments not used in the training and the ability to correctly infer cell type or state by querying a database of tens of thousands of single cell profiles. Such database queries (which can be performed using our web server) will enable researchers to better characterize cells when analyzing heterogeneous scRNA-Seq samples.

Alveolar epithelial type 2 cells (AEC2s) are the facultative progenitors responsible for maintaining lung alveoli throughout life but are difficult to isolate from patients. Here, we engineer AEC2s from human pluripotent stem cells (PSCs) in vitro and use timeseries single-cell RNA sequencing with lentiviral barcoding to profile the kinetics of their differentiation in comparison to primary fetal and adult AEC2 benchmarks. We observe bifurcating cell-fate trajectories as primordial lung progenitors differentiate in vitro, with some progeny reaching their AEC2 fate target, while others diverge to alternative non-lung endodermal fates. We develop a Continuous State Hidden Markov model to identify the timing and type of signals, such as overexuberant Wnt responses, that induce some early multipotent NKX2-1 + progenitors to lose lung fate. Finally, we find that this initial developmental plasticity is regulatable and subsides over time, ultimately resulting in PSC-derived AEC2s that exhibit a stable phenotype and nearly limitless selfrenewal capacity.

Recent studies combine two novel technologies, single-cell RNA-sequencing and CRISPR-Cas9 barcode editing for elucidating developmental lineages at the whole organism level. While these studies provided several insights, they face several computational challenges. First, lineages are reconstructed based on noisy and often saturated random mutation data. Additionally, due to the randomness of the mutations, lineages from multiple experiments cannot be combined to reconstruct a species-invariant lineage tree. To address these issues we developed a statistical method, LinTIMaT, which reconstructs cell lineages using a maximum-likelihood framework by integrating mutation and expression data. Our analysis shows that expression data helps resolve the ambiguities arising in when lineages are inferred based on mutations alone, while also enabling the integration of different individual lineages for the reconstruction of an invariant lineage tree. LinTIMaT lineages have better cell type coherence, improve the functional significance of gene sets and provide new insights on progenitors and differentiation pathways.

The complementary characteristics of active and passive depth sensing techniques motivate the fusion of the Li-DAR sensor and stereo camera for improved depth perception. Instead of directly fusing estimated depths across LiDAR and stereo modalities, we take advantages of the stereo matching network with two enhanced techniques: Input Fusion and Conditional Cost Volume Normalization (CCVNorm) on the LiDAR information. The proposed framework is generic and closely integrated with the cost volume component that is commonly utilized in stereo matching neural networks. We experimentally verify the efficacy and robustness of our method on the KITTI Stereo and Depth Completion datasets, obtaining favorable performance against various fusion strategies. Moreover, we demonstrate that, with a hierarchical extension of CCVNorm, the proposed method brings only slight overhead to the stereo matching network in terms of computation time and model size.

Motivation Methods for reconstructing developmental trajectories from time-series single-cell RNA-Seq (scRNA-Seq) data can be largely divided into two categories. The first, often referred to as pseudotime ordering methods are deterministic and rely on dimensionality reduction followed by an ordering step. The second learns a probabilistic branching model to represent the developmental process. While both types have been successful, each suffers from shortcomings that can impact their accuracy. Results We developed a new method based on continuous-state HMMs (CSHMMs) for representing and modeling time-series scRNA-Seq data. We define the CSHMM model and provide efficient learning and inference algorithms which allow the method to determine both the structure of the branching process and the assignment of cells to these branches. Analyzing several developmental single-cell datasets, we show that the CSHMM method accurately infers branching topology and correctly and continuously assign cells to paths, improving upon prior methods proposed for this task. Analysis of genes based on the continuous cell assignment identifies known and novel markers for different cell types. Availability and implementation Software and Supporting website: www.andrew.cmu.edu/user/chiehl1/CSHMM/ Supplementary information Supplementary data are available at Bioinformatics online.