Conversational AI and equity through assessing GPT-3's communication with diverse social groups on contentious topics.

Autoregressive language models, which use deep learning to produce human-like texts, have surged in prevalence. Despite advances in these models, concerns arise about their equity across diverse populations. While AI fairness is discussed widely, metrics to measure equity in dialogue systems are lacking. This paper presents a framework, rooted in deliberative democracy and science communication studies, to evaluate equity in human-AI communication. Using it, we conducted an algorithm auditing study to examine how GPT-3 responded to different populations who vary in sociodemographic backgrounds and viewpoints on crucial science and social issues: climate change and the Black Lives Matter (BLM) movement. We analyzed 20,000 dialogues with 3290 participants differing in gender, race, education, and opinions. We found a substantively worse user experience among the opinion minority groups (e.g., climate deniers, racists) and the education minority groups; however, these groups changed attitudes toward supporting BLM and climate change efforts much more compared to other social groups after the chat. GPT-3 used more negative expressions when responding to the education and opinion minority groups. We discuss the social-technological implications of our findings for a conversational AI system that centralizes diversity, equity, and inclusion.

GET: a foundation model of transcription across human cell types.

Transcriptional regulation, involving the complex interplay between regulatory sequences and proteins, directs all biological processes. Computational models of transcription lack generalizability to accurately extrapolate in unseen cell types and conditions. Here, we introduce GET, an interpretable foundation model designed to uncover regulatory grammars across 213 human fetal and adult cell types. Relying exclusively on chromatin accessibility data and sequence information, GET achieves experimental-level accuracy in predicting gene expression even in previously unseen cell types. GET showcases remarkable adaptability across new sequencing platforms and assays, enabling regulatory inference across a broad range of cell types and conditions, and uncovering universal and cell type specific transcription factor interaction networks. We evaluated its performance on prediction of regulatory activity, inference of regulatory elements and regulators, and identification of physical interactions between transcription factors. Specifically, we show GET outperforms current models in predicting lentivirus-based massive parallel reporter assay readout with reduced input data. In fetal erythroblasts, we identify distal (>1Mbp) regulatory regions that were missed by previous models. In B cells, we identified a lymphocyte-specific transcription factor-transcription factor interaction that explains the functional significance of a leukemia-risk predisposing germline mutation. In sum, we provide a generalizable and accurate model for transcription together with catalogs of gene regulation and transcription factor interactions, all with cell type specificity.

The immunoregulatory protein CD200 as a potentially lucrative yet elusive target for cancer therapy.

CD200 is an immunoregulatory cell surface ligand with proven pro-tumorigenic credentials via its ability to suppress CD200 receptor (CD200R)-expressing anti-tumor immune function. This definitive role for the CD200-CD200R axis in regulating an immunosuppressive tumor microenvironment has garnered increasing interest in CD200 as a candidate target for immune checkpoint inhibition therapy. However, while the CD200 blocking antibody samalizumab is still in the early stages of clinical testing, alternative mechanisms for the pro-tumorigenic role of CD200 have recently emerged that extend beyond direct suppression of anti-tumor T cell responses and, as such, may not be susceptible to CD200 antibody blockade. Herein, we will summarize the current understanding of CD200 expression and function in the tumor microenvironment as well as alternative strategies for potential neutralization of multiple CD200 mechanisms in human cancers.

The club convergence of green productivity across African countries.

This study investigates economic convergence and sustainable development in Africa. By introducing an aggregate production technology and directional distance function, it examines the productivity growth of 28 African economies from 1990 to 2019. The proposed approach considers all decision-making units (countries) as a whole, and the productivity gains are then estimated under a nonparametric framework. In the empirical analysis, the carbon emissions are included in the Luenberger productivity measurement, called green productivity. The results show that the annual average growth rate of green productivity is 1.51% in African, and different types of club convergence for green productivity indicator and its decomposition are observed during the sample period. The decomposition of the Luenberger indicator shows that green African growth is mainly driven by technological progress, not efficiency change. Furthermore, the overall inefficiency is decomposed into technical and structural effects. The latter measure the potential improvement in terms of resource reallocation. Structural inefficiency is larger than technical inefficiency, suggesting that African countries could improve their economic and environmental performances by optimizing input/output mixes.

The CD200-CD200R Axis Promotes Squamous Cell Carcinoma Metastasis via Regulation of Cathepsin K.

The CD200-CD200R immunoregulatory signaling axis plays an etiologic role in the survival and spread of numerous cancers, primarily through suppression of antitumor immune surveillance. Our previous work outlined a prometastatic role for the CD200-CD200R axis in cutaneous squamous cell carcinoma (cSCC) that is independent of direct T-cell suppression but modulates the function of infiltrating myeloid cells. To identify effectors of the CD200-CD200R axis important for cSCC metastasis, we conducted RNA sequencing profiling of infiltrating CD11B+Cd200R+ cells isolated from CD200+ versus CD200-null cSCCs and identified the cysteine protease cathepsin K (Ctsk) to be highly upregulated in CD200+ cSCCs. CD11B+Cd200R+ cells expressed phenotypic markers associated with myeloid-derived suppressor cell-like cells and tumor-associated macrophages and were the primary source of Ctsk expression in cSCC. A Cd200R+ myeloid cell-cSCC coculture system showed that induction of Ctsk was dependent on engagement of the CD200-CD200R axis, indicating that Ctsk is a target gene of this pathway in the cSCC tumor microenvironment. Inhibition of Ctsk, but not matrix metalloproteinases, significantly blocked cSCC cell migration in vitro. Finally, targeted CD200 disruption in tumor cells and Ctsk pharmacologic inhibition significantly reduced cSCC metastasis in vivo. Collectively, these findings support the conclusion that CD200 stimulates cSCC invasion and metastasis via induction of Ctsk in CD200R+ infiltrating myeloid cells. SIGNIFICANCE: These findings highlight the relationship between CD200-CD200R and cathepsin K in cutaneous squamous cell carcinoma metastasis and suggest that either of these components may serve as a viable therapeutic target in this disease.

Reliable or not? An automated classification of webpages about early childhood vaccination using supervised machine learning.

OBJECTIVE: To investigate the applicability of supervised machine learning (SML) to classify health-related webpages as 'reliable' or 'unreliable' in an automated way. METHODS: We collected the textual content of 468 different Dutch webpages about early childhood vaccination. Webpages were manually coded as 'reliable' or 'unreliable' based on their alignment with evidence-based vaccination guidelines. Four SML models were trained on part of the data, whereas the remaining data was used for model testing. RESULTS: All models appeared to be successful in the automated identification of unreliable (F1 scores: 0.54-0.86) and reliable information (F1 scores: 0.82-0.91). Typical words for unreliable information are 'dr', 'immune system', and 'vaccine damage', whereas 'measles', 'child', and 'immunization rate', were frequent in reliable information. Our best performing model was also successful in terms of out-of-sample prediction, tested on a dataset about HPV vaccination. CONCLUSION: Automated classification of online content in terms of reliability, using basic classifiers, performs well and is particularly useful to identify reliable information. PRACTICE IMPLICATIONS: The classifiers can be used as a starting point to develop more complex classifiers, but also warning tools which can help people evaluate the content they encounter online.

Temperature elevation and Vibrio cyclitrophicus infection reduce the diversity of haemolymph microbiome of the mussel Mytilus coruscus.

Haemolymph microbiome was considered to be unique to healthy invertebrates and beneficial to the host against external pathogens, including disease resistance and maintenance of homeostasis. Here, we investigated the effects of elevated water temperature on infection of haemolymph microbiome of the hard-shelled mussel (Mytilus coruscus). Exposure to Vibrio. cyclitrophicus resulted in high mortality of mussels on day nine at 27 °C. The haemolymph was collected to determine the microbiota by 16 S rRNA gene sequencing. Exposure to waterborne V. cyclitrophicus increased the mortality of mussels that was associated with a reduction in the diversity of their microbial community. Principal coordinate analysis (PCoA) revealed that temperature was an essential factor in shaping microbial communities in mussel haemolymph. Vibrio exposure promoted the proliferation of opportunistic pathogens (e.g., Arcobacter and Francisella) at a lower temperature. A high abundance of Vibrio present in live and dead mussels, at 27 °C might contribute greatly to mortality, as indicated by linear discriminant analysis effect size (LEfSe). These data suggested that the dynamics of microbial community have unique biomarker species in mussel haemolymph that could be used as health indicators. An elevated temperature may reduce the ability of bacterial elimination function against infection in mussel haemolymph.

Characterization of Gut Microbiome in the Mussel Mytilus galloprovincialis in Response to Thermal Stress.

The gut microbiota is essential for utilization of energy and nutrition and may have a role in host immunity in response to environmental shifts. The present study evaluated the temperature stress (increasing from 21 to 27°C) on gut microbiome and dynamics of the mussel Mytilus galloprovincialis by 16S rRNA gene sequencing with the aim of discovering the gut microbiome resilience to warming. Exposure to high temperature of 27°C significantly reduced the survival of M. galloprovincialis associated with increased microbial diversity of gut. The microbial communities were shifted with elevated temperature (from 21 to 27°C) and different exposure time (from day 0 to day 7) by principal coordinate analysis (PCoA). Linear discriminant analysis effect size (LEfSe) revealed that the relative abundance of Vibrio and Arcobacter presented in live animals as the top genus-level biomarkers during the initial exposure to 27°C and followed by microbiomes fluctuation with increasing exposure time at day 4 and day 7. The proliferation of opportunistic pathogens such as genus Vibrio and Arcobacter might increase host susceptibility to disease and contributed greatly to mortality. The results obtained in this study provide the knowledge on ecological adaptation for south domestication of M. galloprovincialis and host-bacteria interaction during temperature stress (27°C).