Endorcine cells producing regulatory peptides

Summary Recent data on the immunologication of regulatory peptides and related propeptide sequences in endocrine cells and tumours of the gastrointestinal tract pancreas, lung, thyroid, pituitary (ACTH and opioids), adrenals and paraganglia have been revised and discussed. Gastrin, xenopsin, cholecystokinin (CCK), somatostatin, motilin, secretin, GIP (gastric inhibitory beenrevised and discussed. Gastrin, xenopsin, cholecystokinin (CCK), somatostatin, motilin, secretin, GIP (gastric inhibitory polypeptide), neurotensin, glicentin/glucagon-37 and PYY (peptide tyrosine tyrosine) are the main products of gastrointestinal endocrine cells; glucagon, CRF (corticotropin releasing factor), somatostatin, PP (pancreatic polypeptide) and GRF (growth hormone releasing factor), in addition to insulin, are produced in pancreatic islet cells; bombesin-related peptidesare the main markers of pulmonary endocrine cells; calcitonin and CGRP (calcitonin gene-related peptide) occur in thyroid and extrathyroid C cells; ACTH and endorphins in anterior and intermediate lobe pituitary cells, -MSH and CLIP (corticotropoin-like intermediate lobe peptide) in intermediate lobe cells; met- and leu-enkephalins and related peptides in adrenal medullary and paraganglionic cells as well as in some gut (enterochromaffin) cells; NPY (neuropeptide Y) in adrenalin-type adrenal medullary cells, etc.. Both tissue-appropriate and tissue-inappropriate regulatory peptides are produced by endocrine tumours, with inappropriate peptides mostly produced by malignant tumours.     

Perennial N2 supersaturation in an Antarctic lake

The dry valleys of southern Victoria Land in Antarctica contain several closed basins in which perennially ice-covered lakes are found. One of the most unusual features of these lakes is the occurrence of high O2 concentrations in the water column; values ranging from slightly more than saturation to more than four times saturation have been reported. Recently, we considered a bulk O2 budget for Lake Hoare, Antarctica, which led us to suggest that biological processes alone were not sufficient to explain the observed elevated oxygen levels. Consequently, there must be a non-biological source of O2. We suggested that this source results from the exclusion of O2 during the freezing of aerated meltstream water at the bottom of the ice cover, and predicted that this physical mechanism should also enhance the other atmospheric gases. Here we report the results of a study which, for the first time, documents the supersaturation of N2 in a lake. Dissolved N2 levels of 145% and 163% were determined from samples taken just below the ice cover and at a depth of 12 m, respectively. The relatively importance of biological and abiological sources is reflected in the ratio of N2 concentration to O2 concentration. In Lake Hoare this ratio was 1.20 at ice/water interface and 1.05 at 12 m; considerably different from the ratio in equilibrium with air (approximately 1.8). Based on these results, we have determined that about half of the net O2 production in the lake is the result of biological processes.     

Photoperiodic summation is temperature-dependent inPyrrhocoris apterus (L.) (Heteroptera)

Summary InPyrrhocoris apterus, a low temperature, 15°C, prevented the termination of diapause by long days and, unexpectedly, also the induction of diapause by short days. Both responses were enabled at a higher temperature, 26°C. In contrast to current concepts, it was proved that the summation of photoperiodic signals was temperature-dependent, since the morphogenetic development was prevented by starvation.     

A major metabolite of Δ1-tetrahydrocannabinol reduces its cataleptic effect in mice

Summary The results described here demonstrate that THC-induced catalepsy in mice can be substantially inhibited by the prior administration of ¹-THC-7-oic acid, the major metabolite of THC in most species including humans. This raises the possibility that the intensity and duration of action of THC may depend to a large degree on the levels of this metabolite at the sites of action.We thank the National Institute on Drug Abuse for supporting this project by grants DA-02043 and DA-02052 and for supplying all of the cannabinoids. One of us (S.B.) is also the recipient of a Research Scientist Award from NIDA. We are grateful to Kristen Carlson and Thomas Honeyman for helpful suggestions in preparing this report.     

Binding of 11-cis retinaldehyde to the partially purified cellular retinaldehyde binding protein from bovine retinal pigment epithelium

Summary 11-cis retinaldehyde binding analysis was performed on a bovine retinal pigment epithelium preparation of cellular retinaldehyde binding protein (CRALBP), whose purity degree was estimated as 75%. Equilibrium binding studies were carried out measuring the replacement of tritium-labeled with unlabeled 11-cis retinaldehyde at 25°C. Analysis of the experimental data both by a direct curve-fitting procedure utilizing a non linear least square regression analysis and by a conventional Scatchard plot revealed a single non-interacting binding site with an apparent equilibrium constant of 0.9×10^–7 M.A binding stoichiometry of approximately 1 mol of 11-cis retinaldehyde/mol of binding protein can be calculated from the experimental data. Competition studies carried out in the presence of unlabeled trans and cis isomers of Vitamin A derivatives confirm the high degree of specificity of the 11-cis retinaldehyde binding.     

基于自注意力机制的新生儿胆道闭锁识别

新生儿胆道闭锁是新生儿常见的致命疾病之一,并且该病在亚洲的发病率高于世界其他地区.新生儿胆道闭锁需要及时发现及时治疗,然而由于缺少专业的儿科医生和辅助诊疗手段,新生儿父母往往不能及时发现而错过了最佳治疗时间.因此,本文开发了一个具有实际应用价值的预诊算法,通过新生儿粪便图片预测新生儿是否患有新生儿胆道闭锁,并提醒新生儿父母及时就诊.为了让算法在应用场景下识别率更高,本文的算法基于一个真实场景下拍摄的新生儿粪图片数据集开发.首先我们设计了一个自注意力网络模型BANet(Biliary Atresia Network),将图片的浅层特征和深层特征相结合,可以得到更好的分类效果.由于拍摄自应用场景下的图片存在过暗和过曝等问题.通过分析数据集的亮度分布,我们设计了一个自动亮度调节算法解决.此外,图片中的阴影也会对识别结果造成干扰,因此我们在训练阶段增加了一种阴影数据增强方式来缓解这一问题.为验证本文提出算法的有效性,本文设计了一个和医生的对比试验.结果证明BANet在四分类的识别率、二分类的识别率、特异性和敏感性等客观评价指标上占有明显优势.本文提出的BANet能够有效利用图片中的颜色、异常点...     

基于词-标签概率的多标签文本分类研究

针对多标签文本分类任务中如何有效地提取文本特征和获取标签之间潜在的相关性问题,提出一种CNN(convolutional neural networks)结合Bi-LSTM (bi-directional long short-term memory)的模型.首先,通过CNN网络和最大池化提取文本的特征;然后,利用训练的Labeled-LDA(labeled latent dirichlet allocation)模型获取所有词与标签之间的词-标签概率信息;接着,使用Bi-LSTM网络和CNN网络提取当前预测文本中每个词的词-标签信息特征;最后,结合提取的文本特征,预测与当前文本相关联的标签集.实验结果表明,使用词-标签概率获取文本中词与标签之间的相关性信息,能够有效提升模型的F1值.