In single-cell RNA sequencing (scRNA-seq), doublets form when two cells are encapsulated into one reaction volume. The existence of doublets, which appear to be-but are not-real cells, is a key confounder in scRNA-seq data analysis. Computational methods have been developed to detect doublets in scRNA-seq data; however, the scRNA-seq field lacks a comprehensive benchmarking of these methods, making it difficult for researchers to choose an appropriate method for specific analyses. We conducted a systematic benchmark study of nine cutting-edge computational doublet-detection methods. Our study included 16 real datasets, which contained experimentally annotated doublets, and 112 realistic synthetic datasets. We compared doublet-detection methods regarding detection accuracy under various experimental settings, impacts on downstream analyses, and computational efficiencies. Our results show that existing methods exhibited diverse performance and distinct advantages in different aspects. Overall, the DoubletFinder method has the best detection accuracy, and the cxds method has the highest computational efficiency. A record of this paper's transparent peer review process is included in the Supplemental Information.

Induced pluripotent stem cell (iPSC)-derived neural cultures from amyotrophic lateral sclerosis (ALS) patients can model disease phenotypes. However, heterogeneity arising from genetic and experimental variability limits their utility, impacting reproducibility and the ability to track cellular origins of pathogenesis. Here, we present methodologies using single-cell RNA sequencing (scRNA-seq) analysis to address these limitations. By repeatedly differentiating and applying scRNA-seq to motor neurons (MNs) from healthy, familial ALS, sporadic ALS, and genome-edited iPSC lines across multiple patients, batches, and platforms, we account for genetic and experimental variability toward identifying unified and reproducible ALS signatures. Combining HOX and developmental gene expression with global clustering, we anatomically classified cells into rostrocaudal, progenitor, and postmitotic identities. By relaxing statistical thresholds, we discovered genes in iPSC-MNs that were concordantly dysregulated in postmortem MNs and yielded predictive ALS markers in other human and mouse models. Our approach thus revealed early, convergent, and MN-resolved signatures of ALS.

Systematic perturbation of cells followed by comprehensive measurements of molecular and phenotypic responses provides informative data resources for constructing computational models of cell biology. Models that generalize well beyond training data can be used to identify combinatorial perturbations of potential therapeutic interest. Major challenges for machine learning on large biological datasets are to find global optima in a complex multidimensional space and mechanistically interpret the solutions. To address these challenges, we introduce a hybrid approach that combines explicit mathematical models of cell dynamics with a machine-learning framework, implemented in TensorFlow. We tested the modeling framework on a perturbation-response dataset of a melanoma cell line after drug treatments. The models can be efficiently trained to describe cellular behavior accurately. Even though completely data driven and independent of prior knowledge, the resulting de novo network models recapitulate some known interactions. The approach is readily applicable to various kinetic models of cell biology. A record of this paper's Transparent Peer Review process is included in the Supplemental Information.

Transcription factors (TFs) bind regulatory DNA to control gene expression, and mutations to either TFs or DNA can alter binding affinities to rewire regulatory networks and drive phenotypic variation. While studies have profiled energetic effects of DNA mutations extensively, we lack similar information for TF variants. Here, we present STAMMP (simultaneous transcription factor affinity measurements via microfluidic protein arrays), a high-throughput microfluidic platform enabling quantitative characterization of hundreds of TF variants simultaneously. Measured affinities for ∼210 mutants of a model yeast TF (Pho4) interacting with 9 oligonucleotides (>1,800 Kds) reveal that many combinations of mutations to poorly conserved TF residues and nucleotides flanking the core binding site alter but preserve physiological binding, providing a mechanism by which combinations of mutations in cis and trans could modulate TF binding to tune occupancies during evolution. Moreover, biochemical double-mutant cycles across the TF-DNA interface reveal molecular mechanisms driving recognition, linking sequence to function. A record of this paper's Transparent Peer Review process is included in the Supplemental Information.

The recent increase of immunopeptidomics data, obtained by mass spectrometry or binding assays, opens up possibilities for investigating endogenous antigen presentation by the highly polymorphic human leukocyte antigen class I (HLA-I) protein. State-of-the-art methods predict with high accuracy presentation by HLA alleles that are well represented in databases at the time of release but have a poorer performance for rarer and less characterized alleles. Here, we introduce a method based on Restricted Boltzmann Machines (RBMs) for prediction of antigens presented on the Major Histocompatibility Complex (MHC) encoded by HLA genes-RBM-MHC. RBM-MHC can be trained on custom and newly available samples with no or a small amount of HLA annotations. RBM-MHC ensures improved predictions for rare alleles and matches state-of-the-art performance for well-characterized alleles while being less data demanding. RBM-MHC is shown to be a flexible and easily interpretable method that can be used as a predictor of cancer neoantigens and viral epitopes, as a tool for feature discovery, and to reconstruct peptide motifs presented on specific HLA molecules.

Compromised protein homeostasis underlies accumulation of plaques and tangles in Alzheimer's disease (AD). To observe protein turnover at early stages of amyloid beta (Aβ) proteotoxicity, we performed pulse-chase proteomics on mouse brains in three genetic models of AD that knock in alleles of amyloid precursor protein (APP) prior to the accumulation of plaques and during disease progression. At initial stages of Aβ accumulation, the turnover of proteins associated with presynaptic terminals is selectively impaired. Presynaptic proteins with impaired turnover, particularly synaptic vesicle (SV)-associated proteins, have elevated levels, misfold in both a plaque-dependent and -independent manner, and interact with APP and Aβ. Concurrent with elevated levels of SV-associated proteins, we found an enlargement of the SV pool as well as enhancement of presynaptic potentiation. Together, our findings reveal that the presynaptic terminal is particularly vulnerable and represents a critical site for manifestation of initial AD etiology. A record of this paper's transparent peer review process is included in the Supplemental Information.