Sajid Ali

Artificial intelligence (AI) is currently being utilized in a wide range of sophisticated applications, but the outcomes of many AI models are challenging to comprehend and trust due to their black-box nature. Usually, it is essential to understand the reasoning behind an AI model’s decision-making. Thus, the need for eXplainable AI (XAI) methods for improving trust in AI models has arisen. XAI has become a popular research subject within the AI field in recent years. Existing survey papers have tackled the concepts of XAI, its general terms, and post-hoc explainability methods but there have not been any reviews that have looked at the assessment methods, available tools, XAI datasets, and other related aspects. Therefore, in this comprehensive study, we provide readers with an overview of the current research and trends in this rapidly emerging area with a case study example. The study starts by explaining the background of XAI, common definitions, and summarizing recently proposed techniques in XAI for supervised machine learning. The review divides XAI techniques into four axes using a hierarchical categorization system: (i) data explainability, (ii) model explainability, (iii) post-hoc explainability, and (iv) assessment of explanations. We also introduce available evaluation metrics as well as open-source packages and datasets with future research directions. Then, the significance of explainability in terms of legal demands, user viewpoints, and application orientation is outlined, termed as XAI concerns. This paper advocates for tailoring explanation content to specific user types. An examination of XAI techniques and evaluation was conducted by looking at 410 critical articles, published between January 2016 and October 2022, in reputed journals and using a wide range of research databases as a source of information. The article is aimed at XAI researchers who are interested in making their AI models more trustworthy, as well as towards researchers from other disciplines who are looking for effective XAI methods to complete tasks with confidence while communicating meaning from data.

BACKGROUND: A single, high priming dose of tremelimumab (anti-cytotoxic T lymphocyte–associated antigen 4) plus durvalumab (anti–programmed cell death ligand-1), an infusion regimen termed STRIDE (Single Tremelimumab Regular Interval Durvalumab), showed encouraging clinical activity and safety in a phase 2 trial of unresectable hepatocellular carcinoma. METHODS: In this global, open-label, phase 3 trial, the majority of the patients we enrolled with unresectable hepatocellular carcinoma and no previous systemic treatment were randomly assigned to receive one of three regimens: tremelimumab (300 mg, one dose) plus durvalumab (1500 mg every 4 weeks; STRIDE), durvalumab (1500 mg every 4 weeks), or sorafenib (400 mg twice daily). The primary objective was overall survival for STRIDE versus sorafenib. Noninferiority for overall survival for durvalumab versus sorafenib was a secondary objective. RESULTS: In total, 1171 patients were randomly assigned to STRIDE (n=393), durvalumab (n=389), or sorafenib (n=389). The median overall survival was 16.43 months (95% confidence interval [CI], 14.16 to 19.58) with STRIDE, 16.56 months (95% CI, 14.06 to 19.12) with durvalumab, and 13.77 months (95% CI, 12.25 to 16.13) with sorafenib. Overall survival at 36 months was 30.7%, 24.7%, and 20.2%, respectively. The overall survival hazard ratio for STRIDE versus sorafenib was 0.78 (96.02% CI, 0.65 to 0.93; P=0.0035). Overall survival with durvalumab monotherapy was noninferior to sorafenib (hazard ratio, 0.86; 95.67% CI, 0.73 to 1.03; noninferiority margin, 1.08). Median progression-free survival was not significantly different among all three groups. Grade 3/4 treatment-emergent adverse events occurred for 50.5% of patients with STRIDE, 37.1% with durvalumab, and 52.4% with sorafenib. CONCLUSIONS: STRIDE significantly improved overall survival versus sorafenib. Durvalumab monotherapy was noninferior to sorafenib for patients with unresectable hepatocellular carcinoma. (Funded by AstraZeneca; ClinicalTrials.gov number, NCT03298451.)

Several environmental stresses, including biotic and abiotic factors, adversely affect the growth and development of crops, thereby lowering their yield. However, abiotic factors, e.g., drought, salinity, cold, heat, ultraviolet radiations (UVr), reactive oxygen species (ROS), trace metals (TM), and soil pH, are extremely destructive and decrease crop yield worldwide. It is expected that more than 50% of crop production losses are due to abiotic stresses. Moreover, these factors are responsible for physiological and biochemical changes in plants. The response of different plant species to such stresses is a complex phenomenon with individual features for several species. In addition, it has been shown that abiotic factors stimulate multi-gene responses by making modifications in the accumulation of the primary and secondary metabolites. Metabolomics is a promising way to interpret biotic and abiotic stress tolerance in plants. The study of metabolic profiling revealed different types of metabolites, e.g., amino acids, carbohydrates, phenols, polyamines, terpenes, etc, which are accumulated in plants. Among all, primary metabolites, such as amino acids, carbohydrates, lipids polyamines, and glycine betaine, are considered the major contributing factors that work as osmolytes and osmoprotectants for plants from various environmental stress factors. In contrast, plant-derived secondary metabolites, e.g., phenolics, terpenoids, and nitrogen-containing compounds (alkaloids), have no direct role in the growth and development of plants. Nevertheless, such metabolites could play a significant role as a defense by protecting plants from biotic factors such as herbivores, insects, and pathogens. In addition, they can enhance the resistance against abiotic factors. Therefore, metabolomics practices are becoming essential and influential in plants by identifying different phytochemicals that are part of the acclimation responses to various stimuli. Hence, an accurate metabolome analysis is important to understand the basics of stress physiology and biochemistry. This review provides insight into the current information related to the impact of biotic and abiotic factors on variations of various sets of metabolite levels and explores how primary and secondary metabolites help plants in response to these stresses.

Isolates of recently spreading races of yellow rust from wheat and triticale in Europe were analysed using virulence phenotypic data of 2605 isolates sampled in 12 countries between 2000 and 2014. A subset of 239 isolates was investigated by microsatellite markers. At least three races of non‐European origin, termed ‘Warrior’, ‘Kranich’ and ‘Triticale aggressive’, were identified in the post‐2011 population. The Warrior race was already present in high frequencies in the first year of detection in most European countries and to a large extent it replaced the pre‐2011 European population. In contrast, the two other exotic races were localized to certain regions and/or crop type. The presence already of at least six multilocus genotypes of the Warrior race and five genotypes of the Kranich race in the first year of detection and across large areas is consistent with a hypothesis of aerial spread from genetically diverse source populations. A comparison with reference isolates sampled from six continents suggested that the Warrior and Kranich races originated from sexually recombining populations in the centre of diversity of the yellow rust fungus in the near‐Himalayan region of Asia. However, the Triticale aggressive race was most similar to populations in the Middle East/Central Asia. The study illustrated the potential role of sexual Puccinia striiformis populations as a reservoir for new races replacing distant clonal populations.

We have investigated whether human lymphoid cells are able to synthesize and secrete human PRL (hPRL) and to express PRL receptors. Metabolic labeling with [35S]methionine and immunoprecipitation of cell extracts from human mononuclear cells (MNC) and a human T lymphocyte cell line with an antiserum against hPRL revealed protein of M(r) 23,000, identical in size to pituitary hPRL. Dilution curves of lymphocyte immunoreactive hPRL were parallel to those obtained with pituitary hPRL in an immunoradiometric assay using two monoclonal antibodies against hPRL. Polymerase chain reaction experiments with primers located in the coding sequence of hPRL showed that the hPRL gene was expressed in MNC. Furthermore, cDNA cloning and sequence analysis indicated the presence of an extra 5' noncoding exon previously described for decidual hPRL. When MNCs were further separated into B cells, T cells, and monocytes, the expression of hPRL appeared to be mainly associated with the T lymphocyte fraction. The hPRL transcript was also detected in thymocytes and in a set of human lymphoid cell lines. Finally, polymerase chain reaction experiments revealed a ubiquitous distribution of PRL receptor gene expression in B cells, T cells, and monocytes. The presence of the receptor for PRL and production of PRL by T lymphocytes suggest a possible autocrine or paracrine effect of PRL in immune cell function.