Among the brainstem raphe nuclei, the dorsal raphe nucleus (DR) contains the greatest number of Pet1-lineage neurons, a predominantly serotonergic group distributed throughout DR subdomains. These neurons collectively regulate diverse physiology and behavior and are often therapeutically targeted to treat affective disorders. Characterizing Pet1 neuron molecular heterogeneity and relating it to anatomy is vital for understanding DR functional organization, with potential to inform therapeutic separability. Here we use high-throughput and DR subdomain-targeted single-cell transcriptomics and intersectional genetic tools to map molecular and anatomical diversity of DR-Pet1 neurons. We describe up to fourteen neuron subtypes, many showing biased cell body distributions across the DR. We further show that P2ry1-Pet1 DR neurons – the most molecularly distinct subtype – possess unique efferent projections and electrophysiological properties. These data complement and extend previous DR characterizations, combining intersectional genetics with multiple transcriptomic modalities to achieve fine-scale molecular and anatomic identification of Pet1 neuron subtypes.

Although fluorescence microscopy is ubiquitous in biomedical research, microscopy methods reporting is inconsistent and perhaps undervalued. We emphasize the importance of appropriate microscopy methods reporting and seek to educate researchers about how microscopy metadata impact data interpretation. We provide comprehensive guidelines and resources to enable accurate reporting for the most common fluorescence light microscopy modalities. We aim to improve microscopy reporting, thus improving the quality, rigor and reproducibility of image-based science.

Brainstem median raphe (MR) neurons expressing the serotonergic regulator gene Pet1 send collateralized projections to forebrain regions to modulate affective, memory-related, and circadian behaviors. Some Pet1 neurons express a surprisingly incomplete battery of serotonin pathway genes, with somata lacking transcripts for tryptophan hydroxylase 2 (Tph2) encoding the rate-limiting enzyme for serotonin [5-hydroxytryptamine (5-HT)] synthesis, but abundant for vesicular glutamate transporter type 3 (Vglut3) encoding a synaptic vesicle-associated glutamate transporter. Genetic fate maps show these nonclassical, putatively glutamatergic Pet1 neurons in the MR arise embryonically from the same progenitor cell compartment-hindbrain rhombomere 2 (r2)-as serotonergic TPH2 1 MR Pet1 neurons. Well established is the distribution of efferents en masse from r2-derived, Pet1-neurons; unknown is the relationship between these efferent targets and the specific constituent source-neuron subgroups identified as r2-Pet1 Tph2-high versus r2-Pet1 Vglut3-high . Using male and female mice, we found r2-Pet1 axonal boutons segregated anatomically largely by serotonin 1 versus VGLUT3 1 identity. The former present in the suprachiasmatic nucleus, paraventricular nucleus of the thalamus, and olfactory bulb; the latter are found in the hippocampus, cortex, and septum. Thus r2-Pet1 Tph2-high and r2-Pet1 Vglut3-high neurons likely regulate distinct brain regions and behaviors. Some r2-Pet1 boutons encased interneuron somata, forming specialized presynaptic "baskets" of VGLUT3 1 or VGLUT3 1 /5-HT 1 identity; this suggests that some r2-Pet1 Vglut3-high neurons may regulate local networks, perhaps with differential kinetics via glutamate versus serotonin signaling. Fibers from other Pet1 neurons (non-r2-derived) were observed in many of these same baskets, suggesting multifaceted regulation. Collectively, these findings inform brain organization and new circuit nodes for therapeutic considerations.

Measuring the phenotypic effect of treatments on cells through imaging assays is an efficient and powerful way of studying cell biology, and requires computational methods for transforming images into quantitative data that highlights phenotypic outcomes. Here, we present an optimized strategy for learning representations of treatment effects from high-throughput imaging data, which follows a causal framework for interpreting results and guiding performance improvements. We use weakly supervised learning (WSL) for modeling associations between images and treatments, and show that it encodes both confounding factors and phenotypic features in the learned representation. To facilitate their separation, we constructed a large training dataset with Cell Painting images from five different sources to maximize experimental diversity, following insights from our causal analysis. Training a WSL model with this dataset successfully improves downstream performance, and produces a reusable convolutional network for image-based profiling, which we call Cell Painting CNN. We conducted a comprehensive evaluation of our strategy on three publicly available Cell Painting datasets, discovering that representations obtained by the Cell Painting CNN can improve performance in downstream analysis up to 25% with respect to classical features, while also being more computationally efficient.