Gomphrena celosioides is a native Brazilian plant found in the State of Mato Grosso do Sul. It is used in folk medicine to treat kidney diseases, skin diseases, infections, rheumatism, gastrointestinal diseases, and respiratory diseases. It is also used as an abortifacient. To evaluate the effects of the ethanolic extract of Gomphrena celosioides (EEGc) on reproductive performance, embryo development, and chromosome stability, Swiss mice were randomly divided into experimental groups (n = 10). The animals in the control group received the vehicle Tween 80–1% in the proportion of 0.1 mL/10 g of body weight orally, from the first to the 18th gestational day. The animals in the treatment groups received the EEGc (100, 1000, and 2000 mg/kg) from the first to the 18th gestational day. The animals underwent evaluations of their reproductive performance and embryofetal development. The results showed that the EEGc did not change the animals’ final weight, weight gain, uterine weight, or net weight gain. The evaluation showed that the absolute and relative organs’ weights did not vary between the different experimental groups. In addition, the EEGc did not change the numbers of implants, live fetuses, dead fetuses, or fetal resorptions. There were no differences in post-operative loss rates, implantations, or resorptions, nor were there differences in fetal viability or sex ratio. The use of the EEGc did not result in different frequencies of malformations. In addition, the EEGc did not alter the frequency of chromosomal damage or frequency of micronuclei. Based on our findings, we considered the extract of Gomphrena celosioides to be safe for use during pregnancy, although some parameters indicated caution in its use.

Molecular docking is indispensable across computer‑aided discovery. However, its conclusions often hinge more on modeling choices than on software brand or nominal score. In this review, we argue that docking should be treated explicitly as conditional modeling whose interpretability depends on structural provenance, ligand‑state definition, search‑space design, and validation under deployment‑relevant conditions. This framework is intended as a practical reference for evaluating docking rigor across both academic and applied workflows. We highlight recurrent failure modes, cross‑target score comparisons under non‑comparable states, over‑read scores as affinities, under‑modeled solvation/flexibility, and uncritical use of predicted structures, and show how AI both exacerbates and mitigates these risks. We then propose best practices for modern validation (self‑docking as necessary but insufficient; cross‑docking, decoys, apo/predicted structures, and out‑of‑distribution tests as essential complements) and offer a concise FAIR reporting checklist enabling reuse and audit. Looking forward, we contend that the most valuable advances are those that improve deployment‑relevant reliability, pose plausibility, enrich screening, and enhance robustness across receptor uncertainty, rather than tool novelty alone. This review reframes docking success from “obtaining a pose and a score” to earning confidence through transparent workflows and evaluation aligned with real use. In this context, the term “AI-driven” does not imply replacing physics-based docking, but rather expanding the workflow landscape in which classical and machine learning approaches coexist. This review, therefore, treats docking as a unified decision framework spanning both paradigms, with emphasis on how validation, generalization, and reproducibility requirements evolve in AI-assisted workflows. Graphical abstract The alternative text for this image may have been generated using AI.

Pancreatic cancer remains a disease that is very difficult to treat. S100 proteins are small calcium binding proteins with diverse intra- and extracellular functions that modulate different aspects of tumorigenesis, including tumor growth and metastasis. High mobility group box 1 (HMGB1) protein is a multifaceted protein that also actively influences the development and progression of tumors. In this study, we investigate the possible correlations, at the transcript level, between S100s and HMGB1 in pancreatic cancer. For this purpose, we calculated Pearson’s correlations between the transcript levels of 13 cancer-related S100 genes and HMGB1 in a cDNA array containing 19 pancreatic cancer tumor samples, and in 8 human pancreatic cancer cell lines. Statistically significant positive correlations were found in 5.5% (5 out of 91) and 37.4% (34 of 91) of the possible S100/S100 or S100/HMGB1 pairs in cells and tumors, respectively. Our data suggest that many S100 proteins crosstalk in pancreatic tumors either with other members of the S100 family, or with HMGB1. These newly observed interdependencies may be used to further the characterization of pancreatic tumors based on S100 and HMGB1 transcription profiles.