Functional Signatures Research Articles

Macrophages excite muscle spindles with glutamate to bolster locomotion.

The stretch reflex is a fundamental component of the motor system that orchestrates the coordinated muscle contractions underlying movement. At the heart of this process lie the muscle spindles (MS), specialized receptors finely attuned to fluctuations in tension within intrafusal muscle fibres. The tension variation in the MS triggers a series of neuronal eventsincluding an initial depolarization of sensory type Ia afferents that subsequently causes the activation of motoneurons within the spinal cord1,2. This neuronal cascade culminates in the execution of muscle contraction, underscoring a presumed closed-loop mechanism between the musculoskeletal and nervous systems. By contrast, here we report the discovery of a new population of macrophages with exclusive molecular and functional signatures within the MS that express the machinery for synthesizing and releasing glutamate. Using mouse intersectional genetics with optogenetics and electrophysiology, we show that activation of MS macrophages (MSMP) drives proprioceptive sensory neuron firing on a millisecond timescale. MSMP activate spinal circuits, motor neurons and muscles by means of a glutamate-dependent mechanism that excites the MS. Furthermore, MSMP respond to neural and muscle activation by increasing the expression of glutaminase, enabling them to convert the uptakenglutamine released by myocytes during muscle contraction into glutamate. Selective silencing or depletion of MSMP in hindlimb muscles disrupted the modulation of the stretch reflex for force generation and sensory feedback correction, impairing locomotor strategies in mice. Our results have identified a new cellular component, the MSMP, that directly regulates neural activity and muscle contraction. The glutamate-mediated signalling of MSMP and their dynamic response to sensory cues introduce a new dimension to our understanding of sensation and motor action, potentially offering innovative therapeutic approaches in conditions that affect sensorimotor function.

Multimodal neural correlates of childhood psychopathology.

Complex structural and functional changes occurring in typical and atypical development necessitate multidimensional approaches to better understand the risk of developing psychopathology. Here, we simultaneously examined structural and functional brain network patterns in relation to dimensions of psychopathology in the Adolescent Brain Cognitive Development dataset. Several components were identified, recapitulating the psychopathology hierarchy, with the general psychopathology (p) factor explaining most covariance with multimodal imaging features, while the internalizing, externalizing, and neurodevelopmental dimensions were each associated with distinct morphological and functional connectivity signatures. Connectivity signatures associated with the p factor and neurodevelopmental dimensions followed the sensory-to-transmodal axis of cortical organization, which is related to the emergence of complex cognition and risk for psychopathology. Results were consistent in two separate data subsamples and robust to variations in analytical parameters. Although model parameters yielded statistically significant brain-behavior associations in unseen data, generalizability of the model was rather limited for all three latent components (r change from within- to out-of-sample statistics: LC1within=0.36, LC1out=0.03; LC2within=0.34, LC2out=0.05; LC3within=0.35, LC3out=0.07). Our findings help in better understanding biological mechanisms underpinning dimensions of psychopathology, and could provide brain-based vulnerability markers.

Spatially Aware Domain Adaptation Enables Cell Type Deconvolution from Multi-Modal Spatially Resolved Transcriptomics.

Spatially Resolved Transcriptomics (SRT) offers unprecedented opportunities to elucidate the cellular arrangements within tissues. Nevertheless, the absence of deconvolution methods that simultaneously model multi-modal features has impeded progress in understanding cellular heterogeneity in spatial contexts. To address this issue, SpaDA is developed, a novel spatially aware domain adaptation method that integrates multi-modal data (i.e., transcriptomics, histological images, and spatial locations) from SRT to accurately estimate the spatial distribution of cell types. SpaDA utilizes a self-expressive variational autoencoder, coupled with deep spatial distribution alignment, to learn and align spatial and graph representations from spatial multi-modal SRT data and single-cell RNA sequencing (scRNA-seq) data. This strategy facilitates the transfer of cell type annotation information across these two similarity graphs, thereby enhancing the prediction accuracy of cell type composition. The results demonstrate that SpaDA surpasses existing methods in cell type deconvolution and the identification of cell types and spatial domains across diverse platforms. Moreover, SpaDA excels in identifying spatially colocalized cell types and key marker genes in regions of low-quality measurements, exemplified by high-resolution mouse cerebellum SRT data. In conclusion, SpaDA offers a powerful and flexible framework for the analysis of multi-modal SRT datasets, advancing the understanding of complex biological systems.

Single-cell multiomics reveals simvastatin inhibits pan-cancer epithelial-mesenchymal transition via the MEK/ERK pathway in XBP1+ mast cells

Distant metastasis is the leading cause of cancer-related mortality, and achieving survival benefits through advancements in systemic therapy remains challenging. Mast cells play a dual role in shaping the tumor microenvironment (TME) and influencing distant metastasis, underscoring the significant research value of targeting mast cells for systemic therapy in advanced cancer. We investigated variations in mast cell infiltration levels in primary and metastatic malignancies using immunocyte infiltration analysis. Mast cell subsets were identified from pan-cancer distant metastasis single-cell sequencing data through dimensionality reduction clustering and cell type annotation, combined with cell trajectory and communication network analyses. A prognostic model was established using WGCNA and 12 machine learning algorithms to identify potential mast cell targets. Drug sensitivity and Mendelian randomization analyses were conducted to select potential drugs targeting mast cells, and their effects on epithelial-mesenchymal transition (EMT) were validated through in vitro experiments, including wound healing, transwell, and western blot assays. Results revealed that activated mast cells show increased infiltration in metastatic tumors, correlating with poor survival duration. XBP1+ mast cells were identified as key components of the inhibitory TME, potentially involved in EMT activation. Simvastatin was identified as a potential drug, reversing EMT induced by XBP1+ mast cells in pan-cancer. Aberrant activation of MEK/ERK signaling in XBP1+ mast cells can stimulate cancer cell EMT by modulating degranulation, while Simvastatin can inhibit EMT by suppressing degranulation.

Estimating uncertainty in reference-based cell type annotation in single-cell genomics.

Estimating uncertainty in reference-based cell type annotation in single-cell genomics.

Potential biomarkers for MCL1 inhibitor sensitivity

MCL1 is an anti-apoptotic member of the BCL2 protein family, and its overexpression is associated with poor prognosis across various cancers. Small molecule inhibitors targeting MCL1 are currently in clinical trials for TNBC and other malignancies. However, one major challenge in the clinical application of MCL1 inhibitors is the inherent or acquired resistance to these drugs. Additionally, there is a lack of predictive biomarkers to identify which tumors will respond to MCL1 inhibition. We identified a four-gene functional signature that promotes MCL1 inhibitor resistance in TNBC cells. This gene signature (GS) can distinguish resistant from sensitive TNBC cell lines. Factors encoded by these four genes promote MCL1 inhibitor resistance at least in part through regulation of the ERK signaling pathway. This mechanism involves the upregulation of BCL2 and the downregulation of BIM, which contribute to the inhibitor resistance. Thus, we have discovered a functional GS that drives MCL1 inhibitor resistance. Currently, the MCL1 inhibitor GS-9716 is in clinical trials for TNBC therapy. If validated in clinical samples, this GS could potentially serve as a predictive biomarker for therapy response and help guide the selection of combination therapies to enhance the effectiveness of MCL1 inhibitors.

Disruption of macroscale functional network organisation in patients with frontotemporal dementia.

Neurodegenerative dementias have a profound impact on higher-order cognitive and behavioural functions. Investigating macroscale functional networks through cortical gradients provides valuable insights into the neurodegenerative dementia process and overall brain function. This approach allows for the exploration of unimodal-multimodal differentiation and the intricate interplay between functional brain networks. We applied cortical gradients mapping to resting-state functional MRI data of patients with frontotemporal dementia (FTD) (behavioural-bvFTD, non-fluent and semantic) and healthy controls. In healthy controls, the principal gradient maximally distinguished sensorimotor from default-mode network (DMN) and the secondary gradient visual from salience network (SN). In all FTD variants, the principal gradient's unimodal-multimodal differentiation was disrupted. The secondary gradient, however, showed widespread disruptions impacting the interactions among all networks specifically in bvFTD, while semantic and non-fluent variants exhibited more focal alterations in limbic and sensorimotor networks. Additionally, the visual network showed responsive and/or compensatory changes in all patients. Importantly, these disruptions extended beyond atrophy distribution and related to symptomatology in patients with bvFTD. In conclusion, optimal brain function requires networks to operate in a segregated yet collaborative manner. In FTD, our findings indicate a collapse and loss of differentiation between networks not solely explained by atrophy. These specific cortical gradients' fingerprints could serve as a functional signature for identifying early changes in neurodegenerative diseases or potential compensatory processes.

Rediscovering publicly available single-cell data with the DISCO platform.

Single-cell RNA sequencing (scRNA-seq) has emerged as the key technique for studying transcriptomics at the single-cell level. In our previous work, we presented the DISCO database (https://www.immunesinglecell.org/) that integrates publicly available human scRNA-seq data. We now introduce an enhanced version of DISCO, which has expanded fourfold to include >100 million cells from >17 thousand samples. It provides uniformly realigned read count tables, curated metadata, integrated tissue and phenotype specific atlases, and harmonized cell type annotations. It also hosts a single-cell enhanced knowledgebase of cell type ontology and gene signatures relating to cell types and phenotypes. Lastly, it offers a suite of tools for data retrieval, integration, annotation, and mapping, allowing users to construct customized atlases and perform integrated analysis with their own data. These tools are also available in a standalone R package for offline analysis.

EPCO-11. BIASES IN HEALTHY BRAIN SINGLE-CELL DATASETS: FINDING THE BEST COMPARISON FOR GLIOBLASTOMA STUDIES

Abstract Tumor cells undergo metabolic alterations that support their proliferation and simultaneously modify immune populations, resulting in an immunosuppressive tumor microenvironment. This study aims to compare the metabolic expression of immune cells within the tumor microenvironment to those in a healthy environment. However, there is a lack of single-cell transcriptomic data for normal brain cells, necessitating an analysis of different methods to compare immune cells in tumors with those in a healthy brain. To address this, we explored cell-to-cell metabolic variability using public single-cell RNA-seq (scRNA-seq) data of glioblastoma (n=338564), and from immune cells in the brains of post-mortem individuals (n=14177), fetuses (n=40000), and epilepsy patients (n=466), as well as from immune blood cells (n=329762). We employed ssGSEA to generate activity scores for 70 key metabolic pathways from KEGG for each individual cell. Known marker genes were utilized to identify normal cell populations (macrophages, T cells, oligodendrocytes) in datasets lacking cell type annotations. Our results revealed that blood cells exhibited high metabolic activity in macrophages but showed moderate activity in other cell types. Immune blood cells also demonstrated high oxidative phosphorylation (OXPHOS) but low fatty acid activity. Fetal brain cells displayed a lack of heterogeneity and showed high fatty acid activity with low OXPHOS activity. Brain cells from post-mortem individuals displayed heterogeneity but exhibited low metabolic activity across nearly all cells and pathways. Additionally, the hierarchy of pathway activity differed from the other datasets. Epilepsy data revealed a metabolic activity pattern similar to tumor-infiltrating immune cells, with comparable activity hierarchy and heterogeneity, although the sample size was limited for detailed comparison. Understanding the differences between types of datasets is important to choose the most appropriate comparison, allowing us to observe the metabolic alterations in immune cells caused by glioblastoma and to develop better immunotherapy strategies.

Knowledge-based inductive bias and domain adaptation for cell type annotation.

Measurement techniques often result in domain gaps among batches of cellular data from a specific modality. The effectiveness of cross-batch annotation methods is influenced by inductive bias, which refers to a set of assumptions that describe the behavior of model predictions. Different annotation methods possess distinct inductive biases, leading to varying degrees of generalizability and interpretability. Given that certain cell types exhibit unique functional patterns, we hypothesize that the inductive biases of cell annotation methods should align with these biological patterns to produce meaningful predictions. In this study, we propose KIDA, Knowledge-based Inductive bias and Domain Adaptation. The knowledge-based inductive bias constrains the prediction rules learned from the reference dataset, composed of multiple batches, to functional patterns relevant to biology, thereby enhancing the generalization of the model to unseen batches. Since the query dataset also contains gaps from multiple batches, KIDA's domain adaptation employs pseudo labels for self-knowledge distillation, effectively narrowing the distribution gap between model predictions and the query dataset. Benchmark experiments demonstrate that KIDA is capable of achieving accurate cross-batch cell type annotation.

Color and Spatial Frequency Provide Functional Signatures of Retinotopic Visual Areas.

Primate vision relies on retinotopically organized cortical parcels defined by representations of hemifield (upper versus lower visual field), eccentricity (fovea versus periphery), and area (V1, V2, V3, V4). Here we test for functional signatures of these organizing principles. We used fMRI to measure responses to gratings varying in spatial frequency, color, and saturation across retinotopically defined parcels in two macaque monkeys, and we developed a Sparse Supervised Embedding (SSE) analysis to identify stimulus features that best distinguish cortical parcels from each other. Constraining the SSE model to distinguish just eccentricity representations of the voxels revealed the expected variation of spatial frequency and S-cone modulation with eccentricity. Constraining the model according to the dorsal-ventral location and retinotopic area of each voxel provided unexpected functional signatures, which we investigated further with standard univariate analyses. Posterior parcels (V1) were distinguished from anterior parcels (V4) by differential responses to chromatic and luminance contrast, especially of low spatial frequency gratings. Meanwhile, ventral parcels were distinguished from dorsal parcels by differential responses to chromatic and luminance contrast, especially of colors that modulate all three cone types. The dorsal-ventral asymmetry not only resembled differences between candidate dorsal and ventral subdivisions of human V4, but also extended to include all retinotopic visual areas, starting in V1 and increasing from V1 to V4. The results provide insight into the functional roles of different retinotopic areas and demonstrate the utility of Sparse Supervised Embedding as a data-driven tool for generating hypotheses about cortical function and behavior.Significance Statement This study demonstrates a new analysis, Sparse Supervised Embedding (SSE), which promises to be useful for visualizing and understanding complex neuroimaging datasets. The paper uses SSE to explore the functional roles of retinotopic visual areas (V1, V2, V3, V4, V3a, MT). The results show that retinotopic areas parcellated by representations for eccentricity and upper/lower visual hemifield have functional signatures, which are defined by unique combinations of responses to color, spatial frequency, and contrast. The functional signatures provide hypotheses for the different roles that the parcels play in vision and help resolve apparent differences between human and macaque visual cortex organization.

Topographic complexity drives trait composition as well as functional and phylogenetic diversity of understory plant communities in microrefugia: new insights for conservation

Topographic complexity supports the maintenance of a high diversity of microhabitats that may act as important ‘safe havens’ – or microrefugia – for biodiversity. Microrefugia are sites with specific environmental conditions that facilitate the persistence of species during environmental changes and exhibit unique eco-evolutionary dynamics. However, our knowledge about how topographic complexity and related eco-evolutionary selective forces influence the functional and phylogenetic signatures of species assemblages in microrefugia is very limited. Although the conceptual framework on the systematic integration of plant functional traits into the study of refugia is well established, more empirical studies on functional trait composition and functional diversity in refugia are urgently needed for more effective conservation. Here we analyzed the distribution of various plant functional traits and phylogenetic patterns in microhabitats (south- and north-facing slopes, and bottoms) of 30 large topographic depressions (i.e. doline microrefugia) and microhabitats of the surrounding plateaus in two distant forested karst regions. We found that plant assemblages in the understory of dolines and their surroundings are characterized by unique functional values and combinations of traits. Doline bottoms had the highest functional diversity among doline microhabitats and supported plant assemblages with considerably different trait compositions from the plateaus. Bottoms also had the highest phylogenetic diversity. These results suggest that topographic complexity in forested dolines has a significant effect on the distribution of plant functional traits in the understory. High functional and phylogenetic diversity in doline bottoms can have important consequences for the long-term survival of plant populations, highlighting that these microhabitats may provide a higher resilience and support an adaptive community-level response to natural and anthropogenic stressors. Understanding mechanisms that drive the survival of species within microrefugia is required to determine the best conservation and management strategies.

Cytometry masked autoencoder: An accurate and interpretable automated immunophenotyper

Single-cell cytometry data are crucial for understanding the role of the immune system in diseases and responses to treatment. However, traditional methods for annotating cytometry data face challenges in scalability, robustness, and accuracy. We propose a cytometry masked autoencoder (cyMAE), which automates immunophenotyping tasks including cell type annotation. The model upholds user-defined cell type definitions, facilitating interpretability and cross-study comparisons. The training of cyMAE has a self-supervised phase, which leverages large amounts of unlabeled data, followed by fine-tuning on specialized tasks using smaller amounts of annotated data. The cost of training a new model is amortized over repeated inferences on new datasets using the same panel. Through validation across multiple studies using the same panel, we demonstrate that cyMAE delivers accurate and interpretable cellular immunophenotyping and improves the prediction of subject-level metadata. This proof of concept marks a significant step forward for large-scale immunology studies.

Single-cell transcriptomics of rectal organoids from individuals with perianal fistulizing Crohn’s disease reveals patient-specific signatures

Perianal fistulizing Crohn’s disease (CD) is a severe gastrointestinal disorder causing extensive mucosal damage with limited treatment options. Severe manifestations of the disease appear at higher rates in non-Europeans but the genetic and cellular mechanisms driving the disease phenotypes remain poorly understood. Herein, we tested whether pathologic determinants in the epithelial stem cell compartment could be detected at the transcript level in rectal organoids derived from a diverse patient population. Rectal organoid and mucosal cells from endoscopic biopsies of each patient having perianal fistulizing CD or no disease controls were prepared for and sequenced at the single cell level. After cell type annotations based on expressed marker genes, samples were analyzed by principal components, for differential transcript expression, cell type proportions, and pathway enrichment. After QC, we produced 77,044 rectal organoid cells (n = 13 patients; 8 CD, 5 controls) with high quality sequences that identified 10 distinct epithelial subtypes, that we compared to 141,367 mucosal epithelial cells (n = 29 patients; 18 CD, 11 controls). Consistent with mucosal epithelial cells, rectal organoids prominently displayed disease signatures represented by the stem and transit amplifying regions of the rectal crypt, including alterations in transcriptional signatures of metabolic, epigenetic, and proliferating pathways. Organoids also retained their gender- and ancestral-specific gene expression signatures. However, they lacked many of the inflammatory signatures observed in epithelial cells from diseased mucosa. Perianal CD patient derived rectal organoids reflect gene expression signatures related to disease, gender, and ancestry, suggesting they harbor inherent properties amenable to further patient-specific, disease-related experimentation.

LnCeCell 2.0: an updated resource for lncRNA-associated ceRNA networks and web tools based on single-cell and spatial transcriptomics sequencing data.

We describe LnCeCell 2.0 (http://bio-bigdata.hrbmu.edu.cn/LnCeCell), an updated resource for lncRNA-associated competing endogenous RNA (ceRNA) networks and web tools based on single-cell and spatial transcriptomics sequencing(stRNA-seq)data. We have updated the LnCeCell 2.0 database with significantly expanded data and improved features, including(i) 257 single-cell RNA sequencingandstRNA-seqdatasets across 86 diseases/phenotypes and 80 human normal tissues, (ii) 836581 cell-specific and spatial spot-specific ceRNA interactions and functional networks for 1002988 cells and 367971 spatial spots, (iii) 15489 experimentally supported lncRNA biomarkers related to disease pathology, diagnosis and treatment, (iv) detailed annotation of cell type, cell state, subcellular and extracellular locations of ceRNAs through manual curationand (v) ceRNA expression profiles and follow-up clinical information of 20326 cancer patients. Further, a panel of 24 flexible tools (including 8 comprehensive and 16 mini-analysis tools) was developed to investigate ceRNA-regulated mechanisms at single-cell/spot resolution. The CeCellTraject tool, for example, illustrates the detailed ceRNA distribution of different cell populations and explores the dynamic change of the ceRNA network along the developmental trajectory. LnCeCell 2.0 will facilitate the study of fine-tuned lncRNA-ceRNA networks with single-cell and spatial spot resolution, helping us to understand the regulatory mechanisms behind complex microbial ecosystems.

Single-cell RNA-seq analysis revealed the stemness of a specific cluster of B cells in acute lymphoblastic leukemia progression.

Childhood acute lymphoblastic leukemia (ALL) is a common pediatric cancer. The heterogeneous characterization of B cells in ALL progression poses new challenges to researchers. We used single-cell sequencing to explore the critical role of B cells in regulating the ALL immune microenvironment. We collected the single cell (sc) RNA-seq data of ALL and health sample from the gene expression omnibus (GEO) database, the "Seurat" and "harmony" R package was used for quality control and scRNA-seq analysis, in which the CellMarker2.0 database was used for cell type annotation. Subsequently, the FindAllMarkers function was used to identify the differentially expressed genes (DEGs) among various cell types and the DAVID database was applied for the biological process of DEGs. Then, the "inferCNV" package was used for copy number variation, regulons and cell communication were performed by SCENIC tool and CellChat package. The role of the target gene in regulating ALL progression was assessed using RT-qPCR, Transwell and scratch healing assays. We identified nine mainly cell clusters after scRNA-seq analysis, in which the B cells had higher infiltration proportion in the ALL samples and were sub-clustered into five cell sub-groups. The B cells 1 is closely associated with cell proliferation and stemness (TNFAIP3 and KDM5B), and the significant CNV of amplification occurred on chr6 and chr21 that supported stemness of B cells1. RXRB is a key transcription factor mediated the proliferation of B cells 1, which in turn suppressed hematopoietic stem cells (HSCs) proliferation and promoted cytotoxic NK/T cells activation through diverse cell communication ways. One of the key regulators of B cells is MYC, which promotes the migration and invasive ability of cell line leukemia cell lines. This study reveals the stemness characteristics of B cells and their critical role in ALL progression, a finding that provides new potential directions for the development of targeted therapies against ALL.

MRNA expression profile and prognostic values of the CDHR family genes in lung adenocarcinoma

Lung adenocarcinoma (LUAD), the predominant subtype of lung cancer, has a high incidence and annual mortality worldwide. Members of the cadherin-related (CDHR) family are associated with many malignant tumor types. However, their role and clinical significance in LUAD have not been clarified. We analyzed the association of CDHRs mRNA expression profiles with prognostic significance, immune infiltration, and potential biological functional signatures in several public databases. We constructed a co-expressed mRNA network, and performed an intrinsic molecular structure and function enrichment analysis. Our results showed that CDHR2 mRNA expression was upregulated in LUAD, whereas CDHR1, CDHR3, CDHR4, and CDHR5 mRNAs were downregulated. Upregulation of CDHR2 mRNA is associated with a poor prognosis in patients with LUAD. Next, the correlation between CDHR family members and immune infiltration was observed. A receiver operating characteristic curve showed that the CDHR family is valuable for diagnosing LUAD. In this study, we found that CDHR2 mRNA expression was upregulated in LUAD. Upregulation of CDHR2 mRNA was associated with a poor prognosis in patients with LUAD.

Functional activity and connectivity signatures of ketamine and lamotrigine during negative emotional processing: a double-blind randomized controlled fMRI study

Ketamine is a highly effective antidepressant (AD) that targets the glutamatergic system and exerts profound effects on brain circuits during negative emotional processing. Interestingly, the effects of ketamine on brain measures are sensitive to modulation by pretreatment with lamotrigine, which inhibits glutamate release. Examining the antagonistic effects of ketamine and lamotrigine on glutamate transmission holds promise to identify effects of ketamine that are mediated through changes in the glutamatergic system. Investigating this modulation in relation to both the acute and sustained effects of ketamine on functional activity and connectivity during negative emotional processing should therefore provide novel insights. 75 healthy subjects were investigated in a double-blind, single-dose, randomized, placebo-controlled, parallel-group study with three treatment conditions (ketamine, lamotrigine pre-treatment, placebo). Participants completed an emotional face viewing task during ketamine infusion and 24 h later. Acute ketamine administration decreased hippocampal and Default Mode Network (DMN) activity and increased fronto-limbic coupling during negative emotional processing. Furthermore, while lamotrigine abolished the ketamine-induced increase in functional connectivity, it had no acute effect on activity. Sustained (24 h later) effects of ketamine were only found for functional activity, with a significant reduction in the posterior DMN. This effect was blocked by pretreatment with lamotrigine. Our results suggest that both the acute increases in fronto-limbic coupling and the delayed decrease in posterior DMN activity, but not the attenuated limbic and DMN recruitment after ketamine, are mediated by altered glutamatergic transmission.

QuAC: Quick Attribute-Centric Type Inference for Python

Python’s dynamic typing facilitates rapid prototyping and underlies its popularity in many domains. However, dynamic typing reduces the power of many static checking and bug-finding tools. Python type annotations can make these tools more useful. Type inference tools aim to reduce developers’ burden of adding them. However, existing type inference tools struggle to support dynamic features, infer correct types (especially container type parameters and non-builtin types), and run in reasonable time. Inspired by Python’s duck typing, where the attributes accessed on Python expressions characterize their implicit interfaces, we propose QuAC (Quick Attribute-Centric Type Inference for Python). At its core, QuAC collects attribute sets for Python expressions and leverages information retrieval techniques to predict classes from these attribute sets. It also recursively predicts container type parameters. We evaluate QuAC’s performance on popular Python projects. Compared to state-of-the-art non-LLM baselines, QuAC predicts types with high accuracy complementary to those predicted by the baselines while not sacrificing coverage. It also demonstrates clear advantages in predicting container type parameters and non-builtin types and reduces run times. Furthermore, QuAC is nearly two orders of magnitude faster than an LLM-based method while covering nearly half of its errorless non-trivial type predictions. It is also significantly more consistent at predicting container type parameters and non-builtin types than the LLM-based method, regardless of whether the project has ground-truth type annotations.

Type Inference Logics

Type inference is essential for statically-typed languages such as OCaml and Haskell. It can be decomposed into two (possibly interleaved) phases: a generator converts programs to constraints; a solver decides whether a constraint is satisfiable. Elaboration, the task of decorating a program with explicit type annotations, can also be structured in this way. Unfortunately, most machine-checked implementations of type inference do not follow this phase-separated, constraint-based approach. Those that do are rarely executable, lack effectful abstractions, and do not include elaboration. To close the gap between common practice in real-world implementations and mechanizations inside proof assistants, we propose an approach that enables modular reasoning about monadic constraint generation in the presence of elaboration. Our approach includes a domain-specific base logic for reasoning about metavariables and a program logic that allows us to reason abstractly about the meaning of constraints. To evaluate it, we report on a machine-checked implementation of our techniques inside the Coq proof assistant. As a case study, we verify both soundness and completeness for three elaborating type inferencers for the simply typed lambda calculus with Booleans. Our results are the first demonstration that type inference algorithms can be verified in the same form as they are implemented in practice: in an imperative style, modularly decomposed into constraint generation and solving, and delivering elaborated terms to the remainder of the compiler chain.