Generative modeling and latent space arithmetics predict single-cell perturbation response across cell types, studies and species, bioRxiv, 2018-11-30
AbstractAccurately modeling cellular response to perturbations is a central goal of computational biology. While such modeling has been proposed based on statistical, mechanistic and machine learning models in specific settings, no generalization of predictions to phenomena absent from training data (‘out-of-sample’) has yet been demonstrated. Here, we present scGen, a model combining variational autoencoders and latent space vector arithmetics for high-dimensional single-cell gene expression data. In benchmarks across a broad range of examples, we show that scGen accurately models dose and infection response of cells across cell types, studies and species. In particular, we demonstrate that scGen learns cell type and species specific response implying that it captures features that distinguish responding from non-responding genes and cells. With the upcoming availability of large-scale atlases of organs in healthy state, we envision scGen to become a tool for experimental design through in silico screening of perturbation response in the context of disease and drug treatment.
biorxiv bioinformatics 0-100-users 2018Intelligible speech synthesis from neural decoding of spoken sentences, bioRxiv, 2018-11-30
The ability to read out, or decode, mental content from brain activity has significant practical and scientific implications. For example, technology that translates cortical activity into speech would be transformative for people unable to communicate as a result of neurological impairment. Decoding speech from neural activity is challenging because speaking requires extremely precise and dynamic control of multiple vocal tract articulators on the order of milliseconds. Here, we designed a neural decoder that explicitly leverages the continuous kinematic and sound representations encoded in cortical activity to generate fluent and intelligible speech. A recurrent neural network first decoded vocal tract physiological signals from direct cortical recordings, and then transformed them to acoustic speech output. Robust decoding performance was achieved with as little as 25 minutes of training data. Naive listeners were able to accurately identify these decoded sentences. Additionally, speech decoding was not only effective for audibly produced speech, but also when participants silently mimed speech. These results advance the development of speech neuroprosthetic technology to restore spoken communication in patients with disabling neurological disorders.
biorxiv neuroscience 200-500-users 2018Sex differences in gene expression in the human fetal brain, bioRxiv, 2018-11-30
ABSTRACTWidespread structural, chemical and molecular differences have been reported between the male and female human brain. Although several neurodevelopmental disorders are more commonly diagnosed in males, little is known regarding sex differences in early human brain development. Here, we used RNA sequencing data from a large collection of human brain samples from the second trimester of gestation (N = 120) to assess sex biases in gene expression within the human fetal brain. In addition to 43 genes (102 Ensembl transcripts) transcribed from the Y-chromosome in males, we detected sex differences in the expression of 2558 autosomal genes (2723 Ensembl transcripts) and 155 genes on the X-chromosome (207 Ensembl transcripts) at a false discovery rate (FDR) < 0.1. Genes exhibiting sex-biased expression in human fetal brain are enriched for high-confidence risk genes for autism and other developmental disorders. Male-biased genes are enriched for expression in neural progenitor cells, whereas female-biased genes are enriched for expression in Cajal-Retzius cells and glia. All gene- and transcript-level data are provided as an online resource (available at <jatsext-link xmlnsxlink=httpwww.w3.org1999xlink ext-link-type=uri xlinkhref=httpfgen.psycm.cf.ac.ukFBSeq1>httpfgen.psycm.cf.ac.ukFBSeq1<jatsext-link>) through which researchers can search, download and visualize data pertaining to sex biases in gene expression during early human brain development.
biorxiv genomics 200-500-users 2018E-ChRPs Engineered Chromatin Remodeling Proteins for Precise Nucleosome Positioning, bioRxiv, 2018-11-29
SummaryRegulation of chromatin structure is essential for controlling the access of DNA to factors that require association with specific DNA sequences. The ability to alter chromatin organization in a targeted manner would provide a mechanism for directly manipulating DNA-dependent processes and should provide a means to study direct consequences of chromatin structural changes. Here we describe the development and validation of engineered chromatin remodeling proteins (E-ChRPs) for inducing programmable changes in nucleosome positioning by design. We demonstrate that E-ChRPs function both in vivo and in vitro to specifically reposition target nucleosomes and entire nucleosomal arrays, and possess the ability to evict native DNA-binding proteins through their action. E-ChRPs can be designed with a range of targeting modalities, including the SpyCatcher and dCas9 moieties, resulting in high versatility and enabling diverse future applications. Thus, engineered chromatin remodeling proteins represent a simple and robust means to probe regulation of DNA-dependent processes in different chromatin contexts.
biorxiv genetics 100-200-users 2018Carbs versus fat does it really matter for maintaining lost weight?, bioRxiv, 2018-11-28
The most read article of 2018 published in The BMJ (httpsdoi.org10.1136bmj.k4583) claimed that restricting dietary carbohydrates offers a metabolic advantage to burn more calories and thereby help patients maintain lost weight. However, analyzing the data according to the original pre-registered statistical plan resulted in no statistically significant effects of diet composition on energy expenditure. The large reported diet effects on energy expenditure calculated using the revised analysis plan depended on data from subjects with excessive amounts of unaccounted energy. Adjusting the data to be commensurate with energy conservation resulted in a diet effect that was less than half the value reported in The BMJ paper. Furthermore, the measured daily average CO2 production rates were not significantly different between the diets and the reported expenditure differences were due to inaccurate calculations based on false assumptions about diet adherence.
biorxiv physiology 200-500-users 2018Long-Term Effects of a Novel Continuous Remote Care Intervention Including Nutritional Ketosis for the Management of Type 2 Diabetes A 2-year Non-randomized Clinical Trial, bioRxiv, 2018-11-28
ABSTRACTOBJECTIVEStudies on long-term sustainability of low-carbohydrate approaches to treat diabetes are limited. We aim to assess the effects of a continuous care intervention (CCI) on retention, glycemic control, weight, body composition, cardiovascular, liver, kidney, thyroid, inflammatory markers, diabetes medication usage and disease outcomes at 2 years in adults with type 2 diabetes (T2D).RESEARCH DESIGN AND METHODSAn open label, non-randomized, controlled study with 262 and 87 participants with T2D were enrolled in the CCI and usual care (UC) groups, respectively.RESULTSSignificant changes from baseline to 2 years in the CCI group included HbA1c (−12% from 7.7±0.1%); fasting glucose (−18% from 163.67±3.90 mgdL); fasting insulin (−42% from 27.73±1.26 pmol L-1); weight (−10% from 114.56±0.60 kg); systolic blood pressure (−4% from 131.7±0.9 mmHg); diastolic blood pressure (−4% from 81.8±0.5 mmHg); triglycerides (−22% from 197.2±9.1 mgdL); HDL-C (+19% from 41.8±0.9 mgdL), and liver alanine transaminase (−21% from 29.16±0.97 UL). Spine bone mineral density in the CCI group was unchanged. Glycemic control medication use (excluding metformin) among CCI participants declined (from 56.9% to 26.8%, P=1.3×10-11) including prescribed insulin (−62%) and sulfonylureas (−100%). The UC group had no significant changes in these parameters (except uric acid and anion gap) or diabetes medication use. There was also significant resolution of diabetes (reversal, 53.5%; remission, 17.6%) in the CCI group but not in UC. All the reported improvements had p-values <0.00012.CONCLUSIONSThe CCI sustained long-term beneficial effects on multiple clinical markers of diabetes and cardiometabolic health at 2 years while utilizing less medication. The intervention was also effective in the resolution of diabetes and visceral obesity, with no adverse effect on bone health.TRIAL REGISTRATIONClinicaltrials.gov <jatsext-link xmlnsxlink=httpwww.w3.org1999xlink ext-link-type=clintrialgov xlinkhref=NCT02519309>NCT02519309<jatsext-link>
biorxiv clinical-trials 200-500-users 2018