LigBuilder: A Multi-Purpose Program for Structure-Based Drug Design

We have developed a new multi-purpose program, LigBuilder, for structure-based drug design. Within the structural constraints of the target protein, LigBuilder builds up ligands step by step using a library of organic fragments. Various operations, such as growing, linking, and mutation, have been implemented to manipulate molecular structures. The user can choose either growing or linking strategies for ligand construction and a genetic algorithm is adopted to control the whole construction process. Binding affinities of the ligands are estimated by an empirical scoring function and the bioavailabilities are evaluated by a set of chemical rules. Using thrombin and dihydrofolate reductase as examples, we have demonstrated that LigBuilder is able to generate chemical structures similar to the known ligands.     

肺癌细胞形态量化和结构特征的抽取

研究了形态十分复杂的肺癌细胞的结构特征,建立了肺癌细胞图像的抽取方法和算法以及相应的ＣＡＮＣＥＲ．ＣＥＬＬ自动分析软件。采用数学形态学的多种逻辑运算和灰度梯度法,并建立高分辨率掩模数字图像扫描方法。该方法可有效地抑制图像中的噪声和干扰,并较好地保持图像的清晰度。应用结果表明所建立的方法提高了肺癌细胞的分析速度和精度,并获得了五种类型肺癌细胞形态的量化结果。为肺癌细胞自动识别的研究提供了一技术手段     

留置导尿生理性自动计量排放器的研制和临床应用

为使截瘫等病人的膀胱仍能象在生理情况下贮存和排放尿液,有效的训练膀胱,并自动准确统计尿量。利用传感、微电脑技术研制出一留置导尿生理性自动计量排放器,初步用于临床50例病人的观察,减少人力,方便护理,效果满意。此仪器还可用于大手术术中术后、及休克等病人的尿量监护。可减少逆行性尿路感染,可推广使用。现报告如下:     

Segmentation of organs-at-risk in cervical cancer CT images with a convolutional neural network

PurposeWe introduced and evaluated an end-to-end organs-at-risk (OARs) segmentation model that can provide accurate and consistent OARs segmentation results in much less time.MethodsWe collected 105 patients’ Computed Tomography (CT) scans that diagnosed locally advanced cervical cancer and treated with radiotherapy in one hospital. Seven organs, including the bladder, bone marrow, left femoral head, right femoral head, rectum, small intestine and spinal cord were defined as OARs. The annotated contours of the OARs previously delineated manually by the patient’s radiotherapy oncologist and confirmed by the professional committee consisted of eight experienced oncologists before the radiotherapy were used as the ground truth masks. A multi-class segmentation model based on U-Net was designed to fulfil the OARs segmentation task. The Dice Similarity Coefficient (DSC) and 95th Hausdorff Distance (HD) are used as quantitative evaluation metrics to evaluate the proposed method.ResultsThe mean DSC values of the proposed method are 0.924, 0.854, 0.906, 0.900, 0.791, 0.833 and 0.827 for the bladder, bone marrow, femoral head left, femoral head right, rectum, small intestine, and spinal cord, respectively. The mean HD values are 5.098, 1.993, 1.390, 1.435, 5.949, 5.281 and 3.269 for the above OARs respectively.ConclusionsOur proposed method can help reduce the inter-observer and intra-observer variability of manual OARs delineation and lessen oncologists’ efforts. The experimental results demonstrate that our model outperforms the benchmark U-Net model and the oncologists’ evaluations show that the segmentation results are highly acceptable to be used in radiation therapy planning.     

Continual improvement of nasopharyngeal carcinoma segmentation with less labeling effort

PurposeConvolutional neural networks (CNNs) offer a promising approach to automated segmentation. However, labeling contours on a large scale is laborious. Here we propose a method to improve segmentation continually with less labeling effort.MethodsThe cohort included 600 patients with nasopharyngeal carcinoma. The proposed method was comprised of four steps. First, an initial CNN model was trained from scratch to perform segmentation of the clinical target volume. Second, a binary classifier was trained using a secondary CNN to identify samples for which the initial model gave a dice similarity coefficient (DSC) < 0.85. Third, the classifier was used to select such samples from the new coming data. Forth, the final model was fine-tuned from the initial model, using only selected samples.ResultsThe classifier can detect poor segmentation of the model with an accuracy of 92%. The proposed segmentation method improved the DSC from 0.82 to 0.86 while reducing the labeling effort by 45%.ConclusionsThe proposed method reduces the amount of labeled training data and improves segmentation by continually acquiring, fine-tuning, and transferring knowledge over long time spans.     

Assessing the fishing footprint using data integrated from different tracking devices: Issues and opportunities

Different tracking devices are currently used to record the activity of fishing vessels in space and time and to analyse the pattern of fishing effort with respect to sea substrates, resources, or socio-economics factors. Among these devices, the Vessel Monitoring System (VMS) and the Automatic Identification System (AIS) represent the most interesting and fertile data sources for fisheries investigations but an integration of these two sources was never explored. Two large VMS and AIS datasets, covering the activity of the Italian fishing vessels with length-over-all ≥15 m for the whole year 2012, were used to perform the first assessment of the features characterizing each data source, and ultimately the properties of the estimators of ecological indicators of fishing pressure obtained using these data. In this way, a dataset related to five vessels equipped by a third high frequency device (the Fishing Observing System) was used as reference for a comparative analysis of fishing activity as reconstructed by VMS and/or AIS. Moreover, the ecological indicator 5 of fishing pressure as spatial extent of fishing activity (defined within the Data Collection Framework for Fisheries—DCF) was used as a reference to test the separate and integrated usage of VMS and AIS data. Results evidenced important differences in spatial, temporal and fleet coverage between VMS and AIS. Namely, VMS is affected by a lower frequency with respect to AIS but covers a larger portion of the fleet. Furthermore, AIS mainly operates near the harbours, so that offshore activity is underrepresented. The integration of VMS and AIS data, however, allows valorising the strengths and minimizing the limits of each source. In fact, the assessment of the total spatial extent of fishing activities (as captured by the DCF indicator 5) is significantly improved by the integrated usage of VMS and AIS data.     

Improved spark and ember detection using stationary wavelet transforms

Calcium sparks and embers are localized intracellular events of calcium release in muscle cells studied frequently by confocal microscopy using line-scan imaging. The large quantity of images and large number of events require automatic detection procedures based on signal processing methods. In the past decades these methods were based on thresholding procedures. Although, recently, wavelet transforms were also introduced, they have not become widespread. We have implemented a set of algorithms based on one- and two-dimensional versions of the à trous wavelet transform. The algorithms were used to perform spike filtering, denoising and detection procedures. Due to the dependence of the algorithms on user adjustable parameters, their effect on the efficiency of the algorithm was studied in detail. We give methods to avoid false positive detections which are the consequence of the background noise in confocal images. In order to establish the efficiency and reliability of the algorithms, various tests were performed on artificial and experimental images. Spark parameters (amplitude, full width-at-half maximum) calculated using the traditional and the wavelet methods were compared. We found that the latter method is capable of identifying more events with better accuracy on experimental images. Furthermore, we extended the wavelet-based transform from calcium sparks to long-lasting small-amplitude events as calcium embers. The method not only solved their automatic detection but enabled the identification of events with small amplitude that otherwise escaped the eye, rendering the determination of their characteristic parameters more accurate.     

Automatic assignment of protein backbone resonances by direct spectrum inspection in targeted acquisition of NMR data

The necessity to acquire large multidimensional datasets, a basis for assignment of NMR resonances, results in long data acquisition times during which substantial degradation of a protein sample might occur. Here we propose a method applicable for such a protein for automatic assignment of backbone resonances by direct inspection of multidimensional NMR spectra. In order to establish an optimal balance between completeness of resonance assignment and losses of cross-peaks due to dynamic processes/degradation of protein, assignment of backbone resonances is set as a stirring criterion for dynamically controlled targeted nonlinear NMR data acquisition. The result is demonstrated with the 12 kDa ¹³C,¹⁵ N-labeled apo-form of heme chaperone protein CcmE, where hydrolytic cleavage of 29 C-terminal amino acids is detected. For this protein, 90 and 98% of manually assignable resonances are automatically assigned within 10 and 40 h of nonlinear sampling of five 3D NMR spectra, respectively, instead of 600 h needed to complete the full time domain grid. In addition, resonances stemming from degradation products are identified. This study indicates that automatic resonance assignment might serve as a guiding criterion for optimal run-time allocation of NMR resources in applications to proteins prone to degradation. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users.     

Segmentation d’images de lésions cérébrales primitives en TEP FET : influence du volume de travail

Purpose(1) Evaluate the reproducibility of segmentation methods depending on the preselection region for tumour volume determination on ¹⁸F-fluoro-ethyl-tyrosine (FET) PET. (2) Evaluate the intra and inter-operator reproducibility of the manual delineation. (3) Compare this delineation with the segmentation methods.Materials and methodsEighteen FET PET of patients with glioblastoma were analysed. Preselection regions were determined prior to any segmentation. Two physicians delineated the tumour volume manually. The tumour volume was also delineated with a threshold method (40 and 70% of SUV_max), and a random walk based method. Pearson coefficient (r) (P < 0.05 for r > 0.468) and Jaccard indices (JI) were used to compare the volumes.ResultsManual delineation was reproducible with r = 0.97 and IJ = 0.65 for intra-operator, and r = 0.76 and IJ = 0.45 for inter-operator reproducibility. The preselection regions for a given lesion were different and the segmentation varied with the preselection region: r = 0.55 JI = 0.58; r = 0.85 JI = 0.83; r = 0.70 JI = 0.39 respectively for the threshold of 40%, 70% and the random walk. The segmentation differed form de manual delineation with r = 0.37 and JI = 0.16; r = 0.54 and JI = 0.42; r = 0.43 and JI = 0.37 respectively for the threshold of 40%, 70% and the random walk.ConclusionThe reproducibility of the segmentation methods depends extensively on the preselection region. The intra-operator reproducibility of cerebral lesion delineation on FET PET is satisfactory. The inter-operator reproducibility could be improved.