Rejoinder to Discussions on “Approval policies for modifications to machine learning‐based software as a medical device: A study of bio‐creep”

We thank the discussants for sharing their unique perspectives on the problem of designing automatic algorithm change protocols (aACPs) for machine learning‐based software as a medical device. Both Pennello et al. and Rose highlighted a number of challenges that arise in real‐world settings, and we whole‐heartedly agree that substantial extensions of our work are needed to understand if and how aACPs can be safely deployed in practice. Our work demonstrated that aACPs that appear to be harmless may allow for biocreep, even when the data distribution is assumed to be representative and stationary over time. While we investigated two solutions that protect against this specific issue, many more statistical and practical challenges remain and we look forward to future research on this topic.     

DNABind: A hybrid algorithm for structure‐based prediction of DNA‐binding residues by combining machine learning‐ and template‐based approaches

Accurate prediction of DNA‐binding residues has become a problem of increasing importance in structural bioinformatics. Here, we presented DNABind, a novel hybrid algorithm for identifying these crucial residues by exploiting the complementarity between machine learning‐ and template‐based methods. Our machine learning‐based method was based on the probabilistic combination of a structure‐based and a sequence‐based predictor, both of which were implemented using support vector machines algorithms. The former included our well‐designed structural features, such as solvent accessibility, local geometry, topological features, and relative positions, which can effectively quantify the difference between DNA‐binding and nonbinding residues. The latter combined evolutionary conservation features with three other sequence attributes. Our template‐based method depended on structural alignment and utilized the template structure from known protein–DNA complexes to infer DNA‐binding residues. We showed that the template method had excellent performance when reliable templates were found for the query proteins but tended to be strongly influenced by the template quality as well as the conformational changes upon DNA binding. In contrast, the machine learning approach yielded better performance when high‐quality templates were not available (about 1/3 cases in our dataset) or the query protein was subject to intensive transformation changes upon DNA binding. Our extensive experiments indicated that the hybrid approach can distinctly improve the performance of the individual methods for both bound and unbound structures. DNABind also significantly outperformed the state‐of‐art algorithms by around 10% in terms of Matthews's correlation coefficient. The proposed methodology could also have wide application in various protein functional site annotations. DNABind is freely available at http://mleg.cse.sc.edu/DNABind/ . Proteins 2013; 81:1885–1899. © 2013 Wiley Periodicals, Inc.     

ITMSQ: A software tool for N‐ and C‐terminal fragment ion pairs based isobaric tandem mass spectrometry quantification

Tandem MS (MS2) quantification using the series of N‐ and C‐terminal fragment ion pairs generated from isobaric‐labelled peptides was recently considered an accurate strategy in quantitative proteomics. However, the presence of multiplexed terminal fragment ion in MS2 spectra may reduce the efficiency of peptide identification, resulting in lower identification scores or even incorrect assignments. To address this issue, we developed a quantitative software tool, denoted isobaric tandem MS quantification (ITMSQ), to improve N‐ and C‐terminal fragment ion pairs based isobaric MS2 quantification. A spectrum splitting module was designed to separate the MS2 spectra from different samples, increasing the accuracy of both identification and quantification. ITMSQ offers a convenient interface through which parameters can be changed along with the labelling method, and the result files and all of the intermediate files can be exported. We performed an analysis of in vivo terminal amino acid labelling labelled HeLa samples and found that the numbers of quantified proteins and peptides increased by 13.64 and 27.52% after spectrum splitting, respectively. In conclusion, ITMSQ provides an accurate and reliable quantitative solutionfor N‐ and C‐terminal fragment ion pairs based isobaric MS2 quantitative methods.     

Smartphone‐based spectral imaging otoscope: System development and preliminary study for evaluation of its potential as a mobile diagnostic tool

We develop a novel smartphone‐based spectral imaging otoscope for telemedicine and examine its capability for the mobile diagnosis of middle ear diseases. The device was applied to perform spectral imaging and analysis of an ear‐mimicking phantom and a normal and abnormal tympanic membrane for evaluation of its potential for the mobile diagnosis. Spectral classified images were obtained via online spectral analysis in a remote server. The phantom experimental results showed that it allowed us to distinguish four different fluids located behind a semitransparent membrane. Also, in the spectral classified images of normal ears (n = 3) and an ear with chronic otitis media (n = 1), the normal and abnormal regions in each ear could be quantitatively distinguished with high contrast. These preliminary results thus suggested that it might have the potentials for providing quantitative information for the mobile diagnosis of various middle ear diseases.     

Combining taxon‐by‐trait and taxon‐by‐site matrices for analysing trait patterns of macroinvertebrate communities: a rejoinder to Monaghan & Soares ()

相似文献   

Dendritic connectivity shapes spatial patterns of genetic diversity: a simulation‐based study

Landscape features notoriously affect spatial patterns of biodiversity. For instance, in dendritic ecological networks (such as river basins), dendritic connectivity has been proposed to create unique spatial patterns of biodiversity. Here, we compared genetic datasets simulated under a lattice‐like, a dendritic and a circular landscape to test the influence of dendritic connectivity on neutral genetic diversity. The circular landscape had a level of connectivity similar to that of the dendritic landscape, so as to isolate the influence of dendricity on genetic diversity. We found that genetic diversity and differentiation varied strikingly among the three landscapes. For instance, the dendritic landscape generated higher total number of alleles and higher global F_st than the lattice‐like landscape, and these indices also varied between the dendritic and the circular landscapes, suggesting an effect of dendricity. Furthermore, in the dendritic landscape, allelic richness was higher in highly connected demes (e.g. confluences in rivers) than in low‐connected demes (e.g. upstream and downstream populations), which was not the case in the circular landscape, hence confirming the major role of dendricity. This led to bell‐shaped distributions of allelic richness along an upstream–downstream gradient. Conversely, genetic differentiation (F_st) was lower in highly than in low‐connected demes (which was not observed in circular landscape), and significant patterns of isolation by distance (IBD) were also observed in the dendritic landscape. We conclude that in dendritic networks, the combined influence of dendricity and connectivity generates unique spatial patterns of neutral genetic diversity, which has implications for population geneticists and conservationists.     

Initial axillary staging of breast cancer using ultrasound‐guided fine needle aspiration: a liquid‐based cytology study

A. Schiettecatte, C. Bourgain, C. Breucq, N. Buls, V. De Wilde and J. de Mey
Initial axillary staging of breast cancer using ultrasound‐guided fine needle aspiration: a liquid‐based cytology study Objective: To evaluate the preoperative detection of axillary metastasis combining ultrasound (US)‐guided fine needle aspiration cytology (FNAC) and liquid‐based cytology (Surepath^®) to reduce sentinel node procedures. Methods: In total, 148 patients with clinically negative lymph nodes and no preoperative therapy were included. All patients underwent preoperative ultrasound of the axilla with FNAC if suspicious lymph nodes were found. Complete axillary lymph node dissection was performed at primary surgery when FNAC was positive. All other patients underwent a sentinel node procedure. Results: US‐guided FNAC of the axilla revealed metastasis in 34 (23.0%) of the 148 patients. These 34 patients were 53.1% of all patients (n = 64) with proven axillary lymph node involvement. In 66 patients (44.6%), both ultrasound and histopathology were negative. Overall sensitivity of US‐guided FNAC was 50.0%, specificity 100%, positive predictive value 100% and negative predictive value 70.2%. In T1 tumours, all patients referred for sentinel node procedure were node‐negative. The correlation between malignant FNAC and histopathology was 100%. US‐guided liquid‐based FNAC in patients with no clinically positive lymph nodes reduced the necessity for a sentinel node procedure by 23.0%. Conclusions: We advocate that US‐guided fine needle aspiration (FNA) combined with liquid‐based cytology of axillary lymph nodes should be included in the preoperative staging of breast cancer.     

Class‐paired Fuzzy SubNETs: A paired variant of the rank‐based network analysis family for feature selection based on protein complexes

Identifying reproducible yet relevant protein features in proteomics data is a major challenge. Analysis at the level of protein complexes can resolve this issue and we have developed a suite of feature‐selection methods collectively referred to as Rank‐Based Network Analysis (RBNA). RBNAs differ in their individual statistical test setup but are similar in the sense that they deploy rank‐defined weights among proteins per sample. This procedure is known as gene fuzzy scoring. Currently, no RBNA exists for paired‐sample scenarios where both control and test tissues originate from the same source (e.g. same patient). It is expected that paired tests, when used appropriately, are more powerful than approaches intended for unpaired samples. We report that the class‐paired RBNA, PPFSNET, dominates in both simulated and real data scenarios. Moreover, for the first time, we explicitly incorporate batch‐effect resistance as an additional evaluation criterion for feature‐selection approaches. Batch effects are class irrelevant variations arising from different handlers or processing times, and can obfuscate analysis. We demonstrate that PPFSNET and an earlier RBNA, PFSNET, are particularly resistant against batch effects, and only select features strongly correlated with class but not batch.     

Seroprevalence of Helicobacter pylori in Hispanics living in Puerto Rico: A population‐based study

Background

Helicobacter pylori is an important etiologic factor for peptic ulcers and gastric cancer, one of the top ten leading causes of cancer death in Puerto Rico. However, the prevalence of H. pylori infections in this population was previously unknown. The aim of this study was to examine the seroprevalence of H. pylori and its associated risk factors in Puerto Rico.

Materials and Methods

A cross‐sectional study was designed using an existing population‐based biorepository. Seropositivity was determined using the Premier^? H. pylori immunoassay. Helicobacter pylori seroprevalence was estimated with 95% confidence using marginal standardization following logistic regression. To assess the risk factors associated with H. pylori seropositivity, a multivariable log‐binomial model was fitted to estimate the prevalence ratio (PR) and its 95% confidence interval (95% CI).

Results

A total of 528 population‐based serum samples were analyzed. The mean age of the study population was 41 ± 12 years, of whom 55.3% were females. The overall seroprevalence of H. pylori was 33.0% (95% CI = 28.3%‐38.1%). Increasing age and having <12 years of education were significantly (P < .05) associated with H. pylori seropositivity in the multivariable model; however, residing in counties with low population density reached marginal significance (P = .085).

Conclusions

We report that H. pylori infection is common among Hispanics living in Puerto Rico. The H. pylori seroprevalence observed in Puerto Rico is similar to the seroprevalence reported in the overall population of the United States. The association between H. pylori seroprevalence and the risk factors analyzed offers insight into the epidemiology of gastric cancer in Puerto Rico and warrants further investigation.

     

The impact of crystallization conditions on structure‐based drug design: A case study on the methylene blue/acetylcholinesterase complex

Structure‐based drug design utilizes apoprotein or complex structures retrieved from the PDB. >57% of crystallographic PDB entries were obtained with polyethylene glycols (PEGs) as precipitant and/or as cryoprotectant, but <6% of these report presence of individual ethyleneglycol oligomers. We report a case in which ethyleneglycol oligomers' presence in a crystal structure markedly affected the bound ligand's position. Specifically, we compared the positions of methylene blue and decamethonium in acetylcholinesterase complexes obtained using isomorphous crystals precipitated with PEG200 or ammonium sulfate. The ligands' positions within the active‐site gorge in complexes obtained using PEG200 are influenced by presence of ethyleneglycol oligomers in both cases bound to W84 at the gorge's bottom, preventing interaction of the ligand's proximal quaternary group with its indole. Consequently, both ligands are ～3.0Å further up the gorge than in complexes obtained using crystals precipitated with ammonium sulfate, in which the quaternary groups make direct π‐cation interactions with the indole. These findings have implications for structure‐based drug design, since data for ligand‐protein complexes with polyethylene glycol as precipitant may not reflect the ligand's position in its absence, and could result in selecting incorrect drug discovery leads. Docking methylene blue into the structure obtained with PEG200, but omitting the ethyleneglycols, yields results agreeing poorly with the crystal structure; excellent agreement is obtained if they are included. Many proteins display features in which precipitants might lodge. It will be important to investigate presence of precipitants in published crystal structures, and whether it has resulted in misinterpreting electron density maps, adversely affecting drug design.     

sandflyDST: a dynamic web‐based decision support tool for the morphological identification of sandflies present in Anatolia and mainland Europe,and user study

Species identification of sandflies is mainly performed according to morphological characters using classical written identification keys. This study introduces a new web‐based decision support tool (sandflyDST) for guiding the morphological identification of sandfly species present in Anatolia and mainland Europe and classified in the Phlebotomus and Sergentomyia genera (both: Diptera: Psychodidae). The current version of the tool consists of 111 questions and 36 drawings obtained from classical written keys, and 107 photographs for the quick and easy identification of 26 species of the genus Phlebotomus and four species of the genus Sergentomyia. The tool guides users through a decision tree using yes/no questions about the morphological characters of the specimen. The tool was applied by 30 individuals, who then completed study questionnaires. The results of subsequent analyses indicated that the usability () and users' level of appreciation (86.6%) of the tool were quite high; almost all of the participants considered recommending the tool to others. The tool may also be useful in training new entomologists and maintaining their level of expertise. This is a dynamic tool and can be improved or upgraded according to feedback. The tool is now available online at http://parasitology.ege.edu.tr/sandflyDST/index.php .     

Modeling G protein‐coupled receptors for structure‐based drug discovery using low‐frequency normal modes for refinement of homology models: Application to H3 antagonists

G Protein‐Coupled Receptors (GPCRs) are integral membrane proteins that play important role in regulating key physiological functions, and are targets of about 50% of all recently launched drugs. High‐resolution experimental structures are available only for very few GPCRs. As a result, structure‐based drug design efforts for GPCRs continue to rely on in silico modeling, which is considered to be an extremely difficult task especially for these receptors. Here, we describe Gmodel, a novel approach for building 3D atomic models of GPCRs using a normal mode‐based refinement of homology models. Gmodel uses a small set of relevant low‐frequency vibrational modes derived from Random Elastic Network model to efficiently sample the large‐scale receptor conformation changes and generate an ensemble of alternative models. These are used to assemble receptor–ligand complexes by docking a known active into each of the alternative models. Each of these is next filtered using restraints derived from known mutation and binding affinity data and is refined in the presence of the active ligand. In this study, Gmodel was applied to generate models of the antagonist form of histamine 3 (H3) receptor. The validity of this novel modeling approach is demonstrated by performing virtual screening (using the refined models) that consistently produces highly enriched hit lists. The models are further validated by analyzing the available SAR related to classical H3 antagonists, and are found to be in good agreement with the available experimental data, thus providing novel insights into the receptor–ligand interactions. Proteins 2010. © 2009 Wiley‐Liss, Inc.     

Tissue rinse liquid‐based cytology: a feasible tool for the intraoperative pathological evaluation of sentinel lymph nodes in breast cancer patients

K. Yamashiro, K. Taira, M. Nakajima, D. Okuyama, M. Azuma, H. Takeda, H. Suzuki, H. Jotoku, K. Watanabe, M. Takahashi, K. Taguchi and M. Tamura
Tissue rinse liquid‐based cytology: a feasible tool for the intraoperative pathological evaluation of sentinel lymph nodes in breast cancer patients Objectives: A unique diagnostic method was designed for the intraoperative pathological evaluation of sentinel lymph nodes (SLNs) in breast cancer patients, and the results were verified with 2 years of experience. Methods: Excised lymph nodes were cut into 2‐mm‐thick slices and rinsed thoroughly in CytoRich Red^®. The sliced tissues were embedded in a paraffin block. Three cytological glass slides of the cells exfoliated in CytoRich Red^® were prepared by the SurePath^® liquid‐based cytology (LBC) technique. Two slides were stained by the Papanicolaou method, and the remaining slide was immunostained with an anti‐keratin antibody. This process is called tissue rinse liquid‐based cytology (TRLBC). The results of TRLBC were compared with those of the final pathological diagnoses, including immunostaining with an anti‐keratin antibody on paraffin blocks (PB). Results: This study analysed 444 SLNs from 247 consecutive breast cancer patients. It required 35 minutes to complete the intraoperative diagnosis on a single node, and it took an additional 5 minutes per node if more than one node was submitted. When the results of PB were assumed to be the gold standard, the sensitivity and specificity of TRLBC were 81.9% and 96.1%, respectively. TRLBC detected all nodes with macrometastasis and 23 of 24 nodes with micrometastasis. Fifteen false‐negative TRLBC results were ‘isolated tumour cell clusters’ on PB, but there was one with micrometastasis histologically. Four of 14 false‐positive TRLBC results were proven to be true positive by supplementary examination using step sectioning of the paraffin blocks of the nodes. Conclusion: TRLBC is a feasible and promising intraoperative cytopathological tool showing a comparable efficacy to PB while still allowing the conventional postoperative histological examination.     

Hybridization selects for prime‐numbered life cycles in Magicicada: An individual‐based simulation model of a structured periodical cicada population

We investigate competition between separate periodical cicada populations each possessing different life‐cycle lengths. We build an individual‐based model to simulate the cicada life cycle and allow random migrations to occur between patches inhabited by the different populations. We show that if hybridization between different cycle lengths produces offspring that have an intermediate life‐cycle length, then predation acts disproportionately to select against the hybrid offspring. This happens because they emerge in low densities without the safety‐in‐numbers provided by either parent population. Thus, prime‐numbered life cycles that can better avoid hybridization are favored. However, we find that this advantage of prime‐numbered cycles occurs only if there is some mechanism that can occasionally synchronize emergence between local populations in sufficiently many patches.     

A review of predation as a limiting factor for bird populations in mesopredator‐rich landscapes: a case study of the UK

The impact of increasing vertebrate predator numbers on bird populations is widely debated among the general public, game managers and conservationists across Europe. However, there are few systematic reviews of whether predation limits the population sizes of European bird species. Views on the impacts of predation are particularly polarised in the UK, probably because the UK has a globally exceptional culture of intensive, high‐yield gamebird management where predator removal is the norm. In addition, most apex predators have been exterminated or much depleted in numbers, contributing to a widely held perception that the UK has high numbers of mesopredators. This has resulted in many high‐quality studies of mesopredator impacts over several decades. Here we present results from a systematic review of predator trends and abundance, and assess whether predation limits the population sizes of 90 bird species in the UK. Our results confirm that the generalist predators Red Fox (Vulpes vulpes) and Crows (Corvus corone and C. cornix) occur at high densities in the UK compared with other European countries. In addition, some avian and mammalian predators have increased numerically in the UK during recent decades. Despite these high and increasing densities of predators, we found little evidence that predation limits populations of pigeons, woodpeckers and passerines, whereas evidence suggests that ground‐nesting seabirds, waders and gamebirds can be limited by predation. Using life‐history characteristics of prey species, we found that mainly long‐lived species with high adult survival and late onset of breeding were limited by predation. Single‐brooded species were also more likely to be limited by predation than multi‐brooded species. Predators that depredate prey species during all life stages (i.e. from nest to adult stages) limited prey numbers more than predators that depredated only specific life stages (e.g. solely during the nest phase). The Red Fox and non‐native mammals (e.g. the American Mink Neovison vison) were frequently identified as numerically limiting their prey species. Our review has identified predator–prey interactions that are particularly likely to result in population declines of prey species. In the short term, traditional predator‐management techniques (e.g. lethal control or fencing to reduce predation by a small number of predator species) could be used to protect these vulnerable species. However, as these techniques are costly and time‐consuming, we advocate that future research should identify land‐use practices and landscape configurations that would reduce predator numbers and predation rates.