Hyb4mC: a hybrid DNA2vec-based model for DNA N4-methylcytosine sites prediction

BACKGROUND

DNA N4-methylcytosine is part of the restrictive modification system, which works by regulating some biological processes, for example, the initiation of DNA replication, mismatch repair and inactivation of transposon. However, using experimental methods to detect 4mC sites is time-consuming and expensive. Besides, considering the huge differences in the number of 4mC samples among different species, it is challenging to achieve a robust multi-species 4mC site prediction performance. Hence, it is of great significance to develop effective computational tools to identify 4mC sites.

RESULTS

This work proposes a flexible deep learning-based framework to predict 4mC sites, called Hyb4mC. Hyb4mC adopts the DNA2vec method for sequence embedding, which captures more efficient and comprehensive information compared with the sequence-based feature method. Then, two different subnets are used for further analysis: Hyb_Caps and Hyb_Conv. Hyb_Caps is composed of a capsule neural network and can generalize from fewer samples. Hyb_Conv combines the attention mechanism with a text convolutional neural network for further feature learning.

CONCLUSIONS

Extensive benchmark tests have shown that Hyb4mC can significantly enhance the performance of predicting 4mC sites compared with the recently proposed methods.

Clinical trial site identification practices and the use of electronic health records in feasibility evaluations: An interview study in the Nordic countries

INTRODUCTION

Feasibility evaluations are performed to create the best possible starting point for the set-up and execution of a clinical trial, and to identify any obstacles for successful trial conduct. New digital technologies can provide various types of data for use in feasibility evaluations. There is a need to identify and compare such data sources for trial site identification and for evaluating the sites' patient recruitment potential. Especially, information is needed on the use of electronic health records. We investigated how different data sources are used by pharmaceutical companies operating in the Nordic countries for identifying trial sites and for evaluating their potential to recruit trial participants.

METHODS

This was a semi-structured qualitative interview study with 21 participants from pharmaceutical companies and contract research organizations operating in Finland, Sweden, Denmark and Norway. Qualitative content analysis was applied.

RESULTS

For identifying countries and trial sites on a global level, the trial sponsors mostly used databases on previous trial performance. The use of electronic health record data was very limited. Sites' and investigators' visibility in various databases was seen as fundamental for their countries becoming selected into new clinical trials. For estimating the sites' recruitment projections, most sites were seen to base their patient count estimates solely on their previous experience. Some sites had reviewed their electronic health record data, which was considered to increase the accuracy of their recruitment estimates and these sites' attractivity. Along with dialogs with investigators, the sponsors used various data sources to validate the investigators' estimates. Legislative obstacles were seen to hinder the use of electronic health record queries for estimation of patient counts.

CONCLUSION

Visibility in the databases used by trial sponsors is crucial for the countries and sites to be identified. Site selection appears to be based on trust and relationships built from experience, but electronic data provide the support upon which the trust is based. Estimation of the number of potential trial participants is a complex and time-consuming process for both investigators and sponsors. Sponsors seem to favour sites who could support their patient count estimates with electronic health record data as they were quicker in providing the estimates and more reliable than sites with no electronic health record evidence. The patient count evaluation process could be simplified, accelerated and made more reliable with more systematic use of electronic health record evidence in the feasibility evaluation phase. This would increase the accuracy of the patient count estimates and, on its part, contribute to improved recruitment success.

Chemical Proteomic Analysis of S-Fatty Acylated Proteins and Their Modification Sites

Protein S-fatty-acylation, the covalent addition of a long-chain fatty acid, predominantly palmitate (S-palmitoylation), to cysteine, is a highly dynamic and regulated process that controls protein function and localization of membrane-associated proteins in eukaryotes. The analysis of S-fatty acylated peptides by mass spectrometry remains challenging due to the hydrophobic and potentially labile thioester linkage of the S-fatty acylated peptides.Here we describe an optimized protocol for the global analysis of S-palmitoylated proteins based on the combination of an alkyne-tagged chemical reporter of palmitoylation, alk-16 with hydroxylamine-selective hydrolysis of thioester bonds. This protocol decreased the number of false positive proteins and was applied to identify S-fatty acylation sites, providing modification sites for 44 proteins out of the 106 S-fatty acylated proteins identified.

A modelling approach for offshore wind farm feasibility with respect to ecosystem-based marine spatial planning

Demand for renewable energy is increasing steadily and regulated by national and international policies. Offshore wind energy sector has been clearly the fastest in its development among other options, and development of new wind farms requires large ocean space. Therefore, there is a need of efficient spatial planning process, including the site selection constrained by technical (wind resource, coastal distance, seafloor) and environmental (impacts) factors and competence of uses. We present a novel approach, using Bayesian Belief Networks (BBN), for an integrated spatially explicit site feasibility identification for offshore wind farms. Our objectives are to: (i) develop a spatially explicit model that integrates the technical, economic, environmental and social dimensions; (ii) operationalize the BBN model; (iii) implement the model at local (Basque Country) and regional (North East Atlantic and Western Mediterranean), and (iv) develop and analyse future scenarios for wind farm installation in a local case study. Results demonstrated a total of 1% (23 km²) of moderate feasibility areas in local scaled analysis, compared to 4% of (21,600 km²) very high, and 5% (30,000 km²) of high feasibility in larger scale analysis. The main challenges were data availability and discretization when trying to expand the model from local to regional level. The use of BBN models to determine the feasibility of offshore wind farm areas has been demonstrated adequate and possible, both at local and regional scales, allowing managers to take management decisions regarding marine spatial planning when including different activities, environmental problems and technological constraints.