Data Quality and Value Sample Clauses

Data Quality and Value. A: The data is not definitive, but rather a wish-list of proteins that modellers would like experimentalists to quantify. We have mapped each protein/gene to: the ID of the protein in the Mouse Genome Informatics Database; ENTREZ ID of mouse gene; ENTREZ ID of human gene; HUGO name of human gene; function (those not yet mapped have been marked NA); reason the protein was included in the list; description/name of the gene/protein; and name of gene, as given in Mouse Genome Informatics Database. B: The generation of the ontology to describe methods was piloted on two papers, and involved in-depth discussions with members of SP5 to make it appropriate and useful. We are confident that by going through this process we have a fairly robust ontology, but refinements will be needed as we curate more papers. C: The KappaNEURON simulator has been verified by comparing reference simulations performed using KappaNEURON and standard NEURON (Sterratt et al 2015). D: The Community Detection suite has been validated against common clustering methods and this evaluation is published (▇▇▇▇▇▇ et al 2016). A: The value of the list of target proteins is to experimentalists, looking to produce the datasets likely to be most use to modellers. B: The ontology and data curation will be of use to modellers searching for parameter values for models. C: The KappaNEURON simulator allows for models of synaptic proteomes with combinatorially large numbers of molecular complexes to be simulated in the context of electrical activity. It is therefore ideally suited for models of synaptic plasticity. D: The Community Detection suite provides, to our knowledge, the only clustering algorithm that performs well without the requirement of iterative fine tuning of parameters on molecular interaction networks of the size, scale and sparseness commonly published.

Data Quality and Value. The quality of the 3D morphological reconstructions was checked systematically by experts in the laboratory.

Data Quality and Value. We have generated reports based on the data pre-processing using the pypreprocess library. These reports will be released with the data. We have made sure that the data are properly aligned and that no large motion (2mm or more) corrupts the acquired images. Then we check a posteriori that the activations obtained after each experiment are in line with our expectations. This posteriori check has remained qualitative so far. Overall, the data have a very high resolution and quality.

Data Quality and Value. The quality of the data found in the literature is controlled by an expert curator. A curator occasionally contacts authors of published work to clarify ambiguous points. UniProt is used every month by ~500,000 users, which highlights the value of the data. Members of the HBP consortium are highly likely to be among these users.

Data Quality and Value. Axon diameter quantification is currently difficult to validate, but PLI provides an independent method to check on reliability. While PLI will be considered as the “ground truth”, MRI will contribute an evaluation of intersubject variability. Further verification is planned, with electromicroscopic data during SGA1 for a few, small regions of interest. Quantification of the microstructure of white matter bundle will have direct applications for simulations, leading for instance to the correct number of axons in each bundle, and for biomarker research in brains of patients with neurological or psychiatric diseases by providing the estimation of the mean and normal range of variability.

Data Quality and Value. Quality of the model is assessed by its capacity to reliably fit the experimental data. Moreover, we check whether the obtained rates are identifiable (not degenerate in the sense that a change in the rate would lead to a worse fit of the data). This exhaustive library of K channel kinetic models is a first. It aims to be a one-stop shop where theoretical neuroscientists will find reliable models taking into account the structural and biophysical properties of the channels. By design, these K channel models are independent of any neuronal data and thus should be appropriate for modelling any type of neurons.