Smoothness without Smoothing: Why Gaussian Naive Bayes Is Not Naive for Multi-Subject Searchlight Studies

Spatial smoothness is helpful when averaging fMRI signals across multiple subjects, as it allows different subjects'' corresponding brain areas to be pooled together even if they are slightly misaligned. However, smoothing is usually not applied when performing multivoxel pattern-based analyses (MVPA), as it runs the risk of blurring away the information that fine-grained spatial patterns contain. It would therefore be desirable, if possible, to carry out pattern-based analyses which take unsmoothed data as their input but which produce smooth images as output. We show here that the Gaussian Naive Bayes (GNB) classifier does precisely this, when it is used in “searchlight” pattern-based analyses. We explain why this occurs, and illustrate the effect in real fMRI data. Moreover, we show that analyses using GNBs produce results at the multi-subject level which are statistically robust, neurally plausible, and which replicate across two independent data sets. By contrast, SVM classifiers applied to the same data do not generate a replication, even if the SVM-derived searchlight maps have smoothing applied to them. An additional advantage of GNB classifiers for searchlight analyses is that they are orders of magnitude faster to compute than more complex alternatives such as SVMs. Collectively, these results suggest that Gaussian Naive Bayes classifiers may be a highly non-naive choice for multi-subject pattern-based fMRI studies.     

Lectin-binding domains on laminin

Chicken erythrocytes have been found to have at least two kinds of phospholipase A2. The first is a soluble enzyme from the cytosole fraction and has no calcium sensitivity. The second can be extracted from the plasma membrane fraction with the nonionic detergent Triton X-100. In this study the membrane-bound enzyme was partially purified by affinity chromatography on phosphatidylcholine-Sepharose, and its specific activity was increased 1100-fold compared with that of the cell homogenate without nuclei. It has an optimum pH of 8.5 and required calcium for maximum activity. It showed the specificity for both phosphatidylcholine and phosphatidylethanolamine, but reacted preferentially on the former substrate. Analysis by concanavalin A-Sepharose affinity chromatography revealed that the membrane-bound phospholipase A2 was retained on the resin and could be eluted specifically with a haptenic sugar, methyl alpha-D-mannopyranoside. The enzyme seems to be either a concanavalin A-binding glycoprotein or a part of a complex with certain concanavalin A-binding glycoproteins.     

A Simple Stimulatory Device for Evoking Point-like Tactile Stimuli: A Searchlight for LFP to Spike Transitions

Current neurophysiological research has the aim to develop methodologies to investigate the signal route from neuron to neuron, namely in the transitions from spikes to Local Field Potentials (LFPs) and from LFPs to spikes.LFPs have a complex dependence on spike activity and their relation is still poorly understood¹. The elucidation of these signal relations would be helpful both for clinical diagnostics (e.g. stimulation paradigms for Deep Brain Stimulation) and for a deeper comprehension of neural coding strategies in normal and pathological conditions (e.g. epilepsy, Parkinson disease, chronic pain). To this aim, one has to solve technical issues related to stimulation devices, stimulation paradigms and computational analyses. Therefore, a custom-made stimulation device was developed in order to deliver stimuli well regulated in space and time that does not incur in mechanical resonance. Subsequently, as an exemplification, a set of reliable LFP-spike relationships was extracted.The performance of the device was investigated by extracellular recordings, jointly spikes and LFP responses to the applied stimuli, from the rat Primary Somatosensory cortex. Then, by means of a multi-objective optimization strategy, a predictive model for spike occurrence based on LFPs was estimated.The application of this paradigm shows that the device is adequately suited to deliver high frequency tactile stimulation, outperforming common piezoelectric actuators. As a proof of the efficacy of the device, the following results were presented: 1) the timing and reliability of LFP responses well match the spike responses, 2) LFPs are sensitive to the stimulation history and capture not only the average response but also the trial-to-trial fluctuations in the spike activity and, finally, 3) by using the LFP signal it is possible to estimate a range of predictive models that capture different aspects of the spike activity.     

Evolutionary perspective on annexin calcium-binding domains

Molecular systematic analysis of the annexin gene superfamily characterized the evolutionary origin, frequency and range of structural variation in calcium interaction domains that are considered intrinsic for membrane targeting and ion channel function. Approximately 36% of annexin repeat domains in an estimated 100 distinct subfamilies contained amino acid changes consistent with the functional loss of type two calcium-binding sites. At least 11% of annexin domains contained a novel K/H/RGD motif conserved in particular subfamilies and manifest in all phyla, apparently via convergent evolution. The first yeast annexin from Yarrowia lipolytica was classified in the ANXC1 subfamily with fungal and mycetozoan representatives. This clade had intact calcium-binding sites but disruption of the normally well-conserved, mid-repeat 4 region implicated in calcium channel regulation. Conversely, a tandem pair of novel annexins from the amphioxus Branchiostoma floridae resembled annexin A13 in gene structure and conserved the charged amino acids associated with the internal hydrophilic pore, but were devoid of external type 2 calcium-binding sites and incorporated K/RGD motifs instead, like annexin A9. The selective erosion of calcium-binding sites in annexin domains and the occurrence of alternate ligands in the same exposed, interhelical loops are pervasive features of the superfamily. This suggests greater complexity than previously appreciated in the mechanisms controlling annexin membrane interaction and calcium channel operation.     

Structural and functional domains on actin

Actin plays several essential roles in cellular processes and is a vital component in the contractile apparatus. To accomplish its many cellular tasks, actin must interact with a wide range of other proteins in addition to self-assembling into filaments. Characterization of these functional domains and localized binding regions on the actin monomer is therefore an important undertaking. Strategies for elucidating the many interaction sites include X-ray diffraction, NMR and fluorescence spectroscopy, chemical modification, chemical cross-linking, protein cleavage, and the study of sequence homologies between the many isotypes of actin. Based on these varied data, we discuss the possible spatial relationships between the interaction sites.     

Low-rank smoothing splines on complicated domains

Smoothing over a domain with irregular boundaries or interior gaps and holes is addressed. Consider the problem of estimating mercury in sediment concentrations in the estuarine waters in New Hampshire. A modified version of low-rank thin plate splines (LTPS) is introduced where the geodesic distance is applied to evaluate dissimilarity of any two data observations: loosely speaking, distances between locations are not measured as the crow flies, but as the fish swims. The method is compared with competing smoothing techniques, LTPS, and finite element L-splines.     

BRCT domains

BRCA1 C-terminal (BRCT) domains are integral signaling modules in the DNA damage response (DDR). Aside from their established roles as phospho-peptide binding modules, BRCT domains have been implicated in phosphorylation-independent protein interactions, DNA binding and poly(ADP-ribose) (PAR) binding. These numerous functions can be attributed to the diversity in BRCT domain structure and architecture, where domains can exist as isolated single domains or assemble into higher order homo- or hetero- domain complexes. In this review, we incorporate recent structural and biochemical studies to demonstrate how structural features allow single and tandem BRCT domains to attain a high degree of functional diversity.     

Ubiquitin-binding domains

Ubiquitin-binding domains (UBDs) are a collection of modular protein domains that non-covalently bind to ubiquitin. These recently discovered motifs interpret and transmit information conferred by protein ubiquitylation to control various cellular events. Detailed molecular structures are known for a number of UBDs, but to understand their mechanism of action, we also need to know how binding specificity is determined, how ubiquitin binding is regulated, and the function of UBDs in the context of full-length proteins. Such knowledge will be key to our understanding of how ubiquitin regulates cellular proteins and processes.     

The C-terminal domains of TARPs: Unexpectedly versatile domains

AMPA receptors mediate the majority of fast synaptic transmission in the central nervous system and are therefore among the most intensively studied ligand-gated ion channels over the last decades. However, the recent discovery that native AMPA receptor complexes contain auxiliary subunits classified as transmembrane AMPA receptor regulatory proteins (TARPs) was quite a surprise and dramatically changed the field of AMPA receptor research. TARPs regulate trafficking as well as synaptic localization of AMPA receptors, and alter their pharmacological and biophysical properties, generally resulting in strongly elevated receptor-mediated currents. Thus, the association of AMPA receptors with TARPs increases receptor heterogeneity and diversity of postsynaptic currents. In this regard, unravelling the mechanisms by which TARPs modulate AMPA receptor function is an intriguing challenge. Studying the functional importance of the carboxy-terminal domain (CTD) of TARPs for receptor modulation, we found that the increased trafficking mediated by the two TARPs γ2 and γ3 is attributable to their CTDs. Furthermore, we demonstrated that the CTD additionally determines the differences between TARPs regarding their modulation of AMPA receptor function. As a case in point, we showed a unique role of the CTD of γ4, suggesting that TARPs modulate AMPA receptor function via individual mechanisms.     

Predicting drug targets based on protein domains

The identification of interactions between drugs and proteins plays key roles in understanding mechanisms underlying drug actions and can lead to new drug design strategies. Here, we present a novel statistical approach, namely PDTD (Predicting Drug Targets with Domains), to predict potential target proteins of new drugs based on derived interactions between drugs and protein domains. The known target proteins of those drugs that have similar therapeutic effects allow us to infer interactions between drugs and protein domains which in turn leads to identification of potential drug-protein interactions. Benchmarking with known drug-protein interactions shows that our proposed methodology outperforms previous methods that exploit either protein sequences or compound structures to predict drug targets, which demonstrates the predictive power of our proposed PDTD method.     

The influence of curvature on membrane domains

An interdependence between local curvature and domain formation has been observed in both cell and model membranes. An implication of this observation is that domain formation in model membranes may be modulated by membrane curvature. In this paper, small-angle neutron scattering (SANS) is used to examine the influence of membrane curvature (i.e., vesicle size) on the formation of membrane domains. It is found that, although vesicle size and polydispersity are not significantly altered by the formation of membrane domains, the area fraction occupied by domains depends on the overall vesicle size. In particular, increasing membrane curvature (i.e., decreasing vesicle size) results in increased area fractions of membrane domains.     

Identification of protein domains on topological basis

A theoretical method is proposed to identify structural domains in proteins of known structures. It is based on the distribution of the local axes of the polypeptide chain. In particular, a statistical analysis is applied to the contributions of the local axes to the absolute writhing number, a topological property of a space curve resulting from the number of self-crossings in the curve projections onto a unit sphere. This finding supports the hypothesis that topological requirements should be satisfied in the process of protein folding and in the final organization of the tertiary structures.     

Membrane binding domains

Eukaryotic signaling and trafficking proteins are rich in modular domains that bind cell membranes. These binding events are tightly regulated in space and time. The structural, biochemical, and biophysical mechanisms for targeting have been worked out for many families of membrane binding domains. This review takes a comparative view of seven major classes of membrane binding domains, the C1, C2, PH, FYVE, PX, ENTH, and BAR domains. These domains use a combination of specific headgroup interactions, hydrophobic membrane penetration, electrostatic surface interactions, and shape complementarity to bind to specific subcellular membranes.     

Protein-promoted membrane domains

The current notion of biological membranes encompasses a very complex structure, made of dynamically changing compartments or domains where different membrane components partition. These domains have been related to important cellular functions such as membrane sorting, signal transduction, membrane fusion, neuronal maturation, and protein activation. Many reviews have dealt with membrane domains where lipid-lipid interactions direct their formation, especially in the case of raft domains, so in this review we considered domains induced by integral membrane proteins. The nature of the interactions involved and the different mechanisms through which membrane proteins segregate lipid domains are presented, in particular with regard to those induced by the nAChR. It may be concluded that coupling of favourable lipid-lipid and lipid-protein interactions is a general condition for this phenomenon to occur.     

Ceramide-enriched membrane domains

Cellular activation involves the re-organization of receptor molecules and the intracellular signalosom in the cell membrane. Recent studies indicate that specialized domains of the cell membrane, termed rafts, are central for the spatial organization of receptors and signaling molecules. Rafts are converted into larger membrane platforms by activity of the acid sphingomyelinase, which hydrolyses raft-sphingomyelin to ceramide. Ceramide molecules spontaneously associate to form ceramide-enriched microdomains, which fuse to large ceramide-enriched membrane platforms. The acid sphingomyelinase is activated by multiple stimuli including CD95, CD40, DR5/TRAIL, CD20, FcgammaRII, CD5, LFA-1, CD28, TNF, the Interleukin-1 receptor, the PAF-receptor, CD14, infection with P. aeruginosa, S. aureus, N. gonorrhoeae, Sindbis-Virus, Rhinovirus, treatment with gamma-irradiation, UV-light, doxorubicin, cisplatin, disruption of integrin-signaling and under some conditions of developmental death. Ceramide-enriched membrane platforms serve the clustering of receptors, the recruitment of intracellular signaling molecules and the exclusion of inhibitory signaling factors and, thus, facilitate signal transduction initiated by the specific stimulus.