Towards generalizable predictions for G protein-coupled receptor variant expression.

Missense mutations that compromise the plasma membrane expression (PME) of integral membrane proteins are the root cause of numerous genetic diseases. Differentiation of this class of mutations from those that specifically modify the activity of the folded protein has proven useful for the development and targeting of precision therapeutics. Nevertheless, it remains challenging to predict the effects of mutations on the stability and/ or expression of membrane proteins. In this work, we utilize deep mutational scanning data to train a series of artificial neural networks to predict the PME of transmembrane domain variants of G protein-coupled receptors from structural and/ or evolutionary features. We show that our best-performing network, which we term the PME predictor, can recapitulate mutagenic trends within rhodopsin and can differentiate pathogenic transmembrane domain variants that cause it to misfold from those that compromise its signaling. This network also generates statistically significant predictions for the relative PME of transmembrane domain variants for another class A G protein-coupled receptor (ß2 adrenergic receptor) but not for an unrelated voltage-gated potassium channel (KCNQ1). Notably, our analyses of these networks suggest structural features alone are generally sufficient to recapitulate the observed mutagenic trends. Moreover, our findings imply that networks trained in this manner may be generalizable to proteins that share a common fold. Implications of our findings for the design of mechanistically specific genetic predictors are discussed.

Introduction to the BioChemical Library (BCL): An Application-Based Open-Source Toolkit for Integrated Cheminformatics and Machine Learning in Computer-Aided Drug Discovery.

The BioChemical Library (BCL) cheminformatics toolkit is an application-based academic open-source software package designed to integrate traditional small molecule cheminformatics tools with machine learning-based quantitative structure-activity/property relationship (QSAR/QSPR) modeling. In this pedagogical article we provide a detailed introduction to core BCL cheminformatics functionality, showing how traditional tasks (e.g., computing chemical properties, estimating druglikeness) can be readily combined with machine learning. In addition, we have included multiple examples covering areas of advanced use, such as reaction-based library design. We anticipate that this manuscript will be a valuable resource for researchers in computer-aided drug discovery looking to integrate modular cheminformatics and machine learning tools into their pipelines.

A Multitask Deep-Learning Method for Predicting Membrane Associations and Secondary Structures of Proteins.

Prediction of residue-level structural attributes and protein-level structural classes helps model protein tertiary structures and understand protein functions. Existing methods are either specialized on only one class of proteins or developed to predict only a specific type of residue-level attribute. In this work, we develop a new deep-learning method, named Membrane Association and Secondary Structure Predictor (MASSP), for accurately predicting both residue-level structural attributes (secondary structure, location, orientation, and topology) and protein-level structural classes (bitopic, α-helical, ß-barrel, and soluble). MASSP integrates a multilayer two-dimensional convolutional neural network (2D-CNN) with a long short-term memory (LSTM) neural network into a multitasking framework. Our comparison shows that MASSP performs equally well or better than the state-of-the-art methods in predicting residue-level secondary structures, boundaries of transmembrane segments, and topology. Furthermore, it achieves outstanding accuracy in predicting protein-level structural classes. MASSP automatically distinguishes the structural classes of input sequences and identifies transmembrane segments and topologies if present, making it broadly applicable to different classes of proteins. In summary, MASSP's good performance and broad applicability make it well suited for annotating residue-level attributes and protein-level structural classes at the proteome scale.

General Purpose Structure-Based Drug Discovery Neural Network Score Functions with Human-Interpretable Pharmacophore Maps.

The BioChemical Library (BCL) is an academic open-source cheminformatics toolkit comprising ligand-based virtual high-throughput screening (vHTS) tools such as quantitative structure-activity/property relationship (QSAR/QSPR) modeling, small molecule flexible alignment, small molecule conformer generation, and more. Here, we expand the capabilities of the BCL to include structure-based virtual screening. We introduce two new score functions, BCL-AffinityNet and BCL-DockANNScore, based on novel distance-dependent signed protein-ligand atomic property correlations. Both metrics are conventional feed-forward dropout neural networks trained on the new descriptors. We demonstrate that BCL-AffinityNet is one of the top performing score functions on the comparative assessment of score functions 2016 affinity prediction and affinity ranking tasks. We also demonstrate that BCL-AffinityNet performs well on the CSAR-NRC HiQ I and II test sets. Furthermore, we demonstrate that BCL-DockANNScore is competitive with multiple state-of-the-art methods on the docking power and screening power tasks. Finally, we show how our models can be decomposed into human-interpretable pharmacophore maps to aid in hit/lead optimization. Altogether, our results expand the utility of the BCL for structure-based scoring to aid small molecule discovery and design. BCL-AffinityNet, BCL-DockANNScore, and the pharmacophore mapping application, as well as the remainder of the BCL cheminformatics toolkit, are freely available with an academic license at the BCL Commons site hosted on http://meilerlab.org/.

BCL::Conf: Improved Open-Source Knowledge-Based Conformation Sampling Using the Crystallography Open Database.

We previously described BCL::Conf, a knowledge-based conformation sampling algorithm utilizing a small molecule fragment rotamer library derived from the Cambridge Structural Database (CSD, license required), as a component of the BioChemical Library (BCL) cheminformatics toolkit. This paper describes substantial improvements made to the BCL::Conf algorithm and a transition to a rotamer library derived from molecules in the Crystallography Open Database (COD, no license required). We demonstrate the performance of the new BCL::Conf on native conformer recovery in the Platinum dataset of high-quality protein-ligand complexes. This set of 2859 structures has previously been used to assess the performance of over a dozen conformer generation algorithms, including the Conformator, Balloon, RDKit DG, ETKDG, Confab, Frog2, MultiConf-DOCK, CSD conformer generator, ConfGenX-OPSL3 force field, Omega, excalc, iCon, and MOE. These benchmarks suggest that the CSD conformer generator is at the state of the art of reported conformer generators. Our results indicate that the improved BCL::Conf significantly outperforms the CSD conformer generation algorithm at binding conformer recovery across a range of ensemble sizes and with similarly fast rates of conformer generation. BCL::Conf is now distributed with the COD-derived rotamer library and is free for academic use. The BCL can be downloaded at http://meilerlab.org/bclcommons for Windows, Linux, or Apple operating systems. BCL::Conf can now also be accessed via webserver at http://meilerlab.org/bclconf.

Efficient Sampling of Protein Loop Regions Using Conformational Hashing Complemented with Random Coordinate Descent.

De novo construction of loop regions is an important problem in computational structural biology. Compared to regions with well-defined secondary structure, loops tend to exhibit significant conformational heterogeneity. As a result, their structures are often ambiguous when determined using experimental data obtained by crystallography, cryo-EM, or NMR. Although structurally diverse models could provide a more relevant representation of proteins in their native states, obtaining large numbers of biophysically realistic and physiologically relevant loop conformations is a resource-consuming task. To address this need, we developed a novel loop construction algorithm, Hash/RCD, that combines knowledge-based conformational hashing with random coordinate descent (RCD). This hybrid approach achieved a closure rate of 100% on a benchmark set of 195 loops in 29 proteins that range from 3 to 31 residues. More importantly, the use of templates allows Hash/RCD to maintain the accuracy of state-of-the-art coordinate descent methods while reducing sampling time from over 400 to 141 ms. These results highlight how the integration of coordinate descent with knowledge-based sampling overcomes barriers inherent to either approach in isolation. This method may facilitate the identification of native-like loop conformations using experimental data or full-atom scoring functions by allowing rapid sampling of large numbers of loops. In this manuscript, we investigate and discuss the advantages, bottlenecks, and limitations of combining conformational hashing with RCD. By providing a detailed technical description of the Hash/RCD algorithm, we hope to facilitate its implementation by other researchers.

Interfaces Between Alpha-helical Integral Membrane Proteins: Characterization, Prediction, and Docking.

Protein-protein interaction (PPI) is an essential mechanism by which proteins perform their biological functions. For globular proteins, the molecular characteristics of such interactions have been well analyzed, and many computational tools are available for predicting PPI sites and constructing structural models of the complex. In contrast, little is known about the molecular features of the interaction between integral membrane proteins (IMPs) and few methods exist for constructing structural models of their complexes. Here, we analyze the interfaces from a non-redundant set of complexes of α-helical IMPs whose structures have been determined to a high resolution. We find that the interface is not significantly different from the rest of the surface in terms of average hydrophobicity. However, the interface is significantly better conserved and, on average, inter-subunit contacting residue pairs correlate more strongly than non-contacting pairs, especially in obligate complexes. We also develop a neural network-based method, with an area under the receiver operating characteristic curve of 0.75 and a Pearson correlation coefficient of 0.70, for predicting interface residues and their weighted contact numbers (WCNs). We further show that predicted interface residues and their WCNs can be used as restraints to reconstruct the structure α-helical IMP dimers through docking for fourteen out of a benchmark set of sixteen complexes. The RMSD100 values of the best-docked ligand subunit to its native structure are <2.5 Šfor these fourteen cases. The structural analysis conducted in this work provides molecular details about the interface between α-helical IMPs and the WCN restraints represent an efficient means to score α-helical IMP docking candidates.

BCL::Mol2D-a robust atom environment descriptor for QSAR modeling and lead optimization.

Comparing fragment based molecular fingerprints of drug-like molecules is one of the most robust and frequently used approaches in computer-assisted drug discovery. Molprint2D, a popular atom environment (AE) descriptor, yielded the best enrichment of active compounds across a diverse set of targets in a recent large-scale study. We present here BCL::Mol2D descriptors that outperformed Molprint2D on nine PubChem datasets spanning a wide range of protein classes. Because BCL::Mol2D records the number of AEs from a universal AE library, a novel aspect of BCL::Mol2D over the Molprint2D is its reversibility. This property enables decomposition of prediction from machine learning models to particular molecular substructures. Artificial neural networks with dropout, when trained on BCL::Mol2D descriptors outperform those trained on Molprint2D descriptors by up to 26% in logAUC metric. When combined with the Reduced Short Range descriptor set, our previously published set of descriptors optimized for QSARs, BCL::Mol2D yields a modest improvement. Finally, we demonstrate how the reversibility of BCL::Mol2D enables visualization of a 'pharmacophore map' that could guide lead optimization for serine/threonine kinase 33 inhibitors.

Protein structure aids predicting functional perturbation of missense variants in SCN5A and KCNQ1.

Rare variants in the cardiac potassium channel KV7.1 (KCNQ1) and sodium channel NaV1.5 (SCN5A) are implicated in genetic disorders of heart rhythm, including congenital long QT and Brugada syndromes (LQTS, BrS), but also occur in reference populations. We previously reported two sets of NaV1.5 (n = 356) and KV7.1 (n = 144) variants with in vitro characterized channel currents gathered from the literature. Here we investigated the ability to predict commonly reported NaV1.5 and KV7.1 variant functional perturbations by leveraging diverse features including variant classifiers PROVEAN, PolyPhen-2, and SIFT; evolutionary rate and BLAST position specific scoring matrices (PSSM); and structure-based features including "functional densities" which is a measure of the density of pathogenic variants near the residue of interest. Structure-based functional densities were the most significant features for predicting NaV1.5 peak current (adj. R2 = 0.27) and KV7.1 + KCNE1 half-maximal voltage of activation (adj. R2 = 0.29). Additionally, use of structure-based functional density values improves loss-of-function classification of SCN5A variants with an ROC-AUC of 0.78 compared with other predictive classifiers (AUC = 0.69; two-sided DeLong test p = .01). These results suggest structural data can inform predictions of the effect of uncharacterized SCN5A and KCNQ1 variants to provide a deeper understanding of their burden on carriers.

BCL::MolAlign: Three-Dimensional Small Molecule Alignment for Pharmacophore Mapping.

Small molecule flexible alignment is a critical component of both ligand- and structure-based methods in computer-aided drug discovery. Despite its importance, the availability of high-quality flexible alignment software packages is limited. Here, we present BCL::MolAlign, a freely available property-based molecular alignment program. BCL::MolAlign accommodates ligand flexibility through a combination of pregenerated conformers and on-the-fly bond rotation. BCL::MolAlign converges on alignment poses by sampling the relative orientations of mutually matching atom pairs between molecules through Monte Carlo Metropolis sampling. Across six diverse ligand data sets, BCL::MolAlign flexible alignment outperforms MOE, ROCS, and FLEXS in recovering native ligand binding poses. Moreover, the BCL::MolAlign alignment score is more predictive of ligand activity than maximum common substructure similarity across 10 data sets. Finally, on a recently published benchmark set of 20 high quality congeneric ligand-protein complexes, BCL::MolAlign is able to recover a larger fraction of native binding poses than maximum common substructure-based alignment and RosettaLigand. BCL::MolAlign can be obtained as part of the Biology and Chemistry Library (BCL) software package freely with an academic license or can be accessed via Web server at http://meilerlab.org/index.php/servers/molalign .

Predicting the Functional Impact of KCNQ1 Variants of Unknown Significance.

BACKGROUND: An emerging standard-of-care for long-QT syndrome uses clinical genetic testing to identify genetic variants of the KCNQ1 potassium channel. However, interpreting results from genetic testing is confounded by the presence of variants of unknown significance for which there is inadequate evidence of pathogenicity. METHODS AND RESULTS: In this study, we curated from the literature a high-quality set of 107 functionally characterized KCNQ1 variants. Based on this data set, we completed a detailed quantitative analysis on the sequence conservation patterns of subdomains of KCNQ1 and the distribution of pathogenic variants therein. We found that conserved subdomains generally are critical for channel function and are enriched with dysfunctional variants. Using this experimentally validated data set, we trained a neural network, designated Q1VarPred, specifically for predicting the functional impact of KCNQ1 variants of unknown significance. The estimated predictive performance of Q1VarPred in terms of Matthew's correlation coefficient and area under the receiver operating characteristic curve were 0.581 and 0.884, respectively, superior to the performance of 8 previous methods tested in parallel. Q1VarPred is publicly available as a web server at http://meilerlab.org/q1varpred. CONCLUSIONS: Although a plethora of tools are available for making pathogenicity predictions over a genome-wide scale, previous tools fail to perform in a robust manner when applied to KCNQ1. The contrasting and favorable results for Q1VarPred suggest a promising approach, where a machine-learning algorithm is tailored to a specific protein target and trained with a functionally validated data set to calibrate informatics tools.

Improving prediction of helix-helix packing in membrane proteins using predicted contact numbers as restraints.

One of the challenging problems in tertiary structure prediction of helical membrane proteins (HMPs) is the determination of rotation of α-helices around the helix normal. Incorrect prediction of helix rotations substantially disrupts native residue-residue contacts while inducing only a relatively small effect on the overall fold. We previously developed a method for predicting residue contact numbers (CNs), which measure the local packing density of residues within the protein tertiary structure. In this study, we tested the idea of incorporating predicted CNs as restraints to guide the sampling of helix rotation. For a benchmark set of 15 HMPs with simple to rather complicated folds, the average contact recovery (CR) of best-sampled models was improved for all targets, the likelihood of sampling models with CR greater than 20% was increased for 13 targets, and the average RMSD100 of best-sampled models was improved for 12 targets. This study demonstrated that explicit incorporation of CNs as restraints improves the prediction of helix-helix packing. Proteins 2017; 85:1212-1221. © 2017 Wiley Periodicals, Inc.

Documentation of an Imperative To Improve Methods for Predicting Membrane Protein Stability.

There is a compelling and growing need to accurately predict the impact of amino acid mutations on protein stability for problems in personalized medicine and other applications. Here the ability of 10 computational tools to accurately predict mutation-induced perturbation of folding stability (ΔΔG) for membrane proteins of known structure was assessed. All methods for predicting ΔΔG values performed significantly worse when applied to membrane proteins than when applied to soluble proteins, yielding estimated concordance, Pearson, and Spearman correlation coefficients of <0.4 for membrane proteins. Rosetta and PROVEAN showed a modest ability to classify mutations as destabilizing (ΔΔG < -0.5 kcal/mol), with a 7 in 10 chance of correctly discriminating a randomly chosen destabilizing variant from a randomly chosen stabilizing variant. However, even this performance is significantly worse than for soluble proteins. This study highlights the need for further development of reliable and reproducible methods for predicting thermodynamic folding stability in membrane proteins.

Improving quantitative structure-activity relationship models using Artificial Neural Networks trained with dropout.

Dropout is an Artificial Neural Network (ANN) training technique that has been shown to improve ANN performance across canonical machine learning (ML) datasets. Quantitative Structure Activity Relationship (QSAR) datasets used to relate chemical structure to biological activity in Ligand-Based Computer-Aided Drug Discovery pose unique challenges for ML techniques, such as heavily biased dataset composition, and relatively large number of descriptors relative to the number of actives. To test the hypothesis that dropout also improves QSAR ANNs, we conduct a benchmark on nine large QSAR datasets. Use of dropout improved both enrichment false positive rate and log-scaled area under the receiver-operating characteristic curve (logAUC) by 22-46 % over conventional ANN implementations. Optimal dropout rates are found to be a function of the signal-to-noise ratio of the descriptor set, and relatively independent of the dataset. Dropout ANNs with 2D and 3D autocorrelation descriptors outperform conventional ANNs as well as optimized fingerprint similarity search methods.

Accurate Prediction of Contact Numbers for Multi-Spanning Helical Membrane Proteins.

Prediction of the three-dimensional (3D) structures of proteins by computational methods is acknowledged as an unsolved problem. Accurate prediction of important structural characteristics such as contact number is expected to accelerate the otherwise slow progress being made in the prediction of 3D structure of proteins. Here, we present a dropout neural network-based method, TMH-Expo, for predicting the contact number of transmembrane helix (TMH) residues from sequence. Neuronal dropout is a strategy where certain neurons of the network are excluded from back-propagation to prevent co-adaptation of hidden-layer neurons. By using neuronal dropout, overfitting was significantly reduced and performance was noticeably improved. For multi-spanning helical membrane proteins, TMH-Expo achieved a remarkable Pearson correlation coefficient of 0.69 between predicted and experimental values and a mean absolute error of only 1.68. In addition, among those membrane protein-membrane protein interface residues, 76.8% were correctly predicted. Mapping of predicted contact numbers onto structures indicates that contact numbers predicted by TMH-Expo reflect the exposure patterns of TMHs and reveal membrane protein-membrane protein interfaces, reinforcing the potential of predicted contact numbers to be used as restraints for 3D structure prediction and protein-protein docking. TMH-Expo can be accessed via a Web server at www.meilerlab.org .

Autocorrelation descriptor improvements for QSAR: 2DA_Sign and 3DA_Sign.

Quantitative structure-activity relationship (QSAR) is a branch of computer aided drug discovery that relates chemical structures to biological activity. Two well established and related QSAR descriptors are two- and three-dimensional autocorrelation (2DA and 3DA). These descriptors encode the relative position of atoms or atom properties by calculating the separation between atom pairs in terms of number of bonds (2DA) or Euclidean distance (3DA). The sums of all values computed for a given small molecule are collected in a histogram. Atom properties can be added with a coefficient that is the product of atom properties for each pair. This procedure can lead to information loss when signed atom properties are considered such as partial charge. For example, the product of two positive charges is indistinguishable from the product of two equivalent negative charges. In this paper, we present variations of 2DA and 3DA called 2DA_Sign and 3DA_Sign that avoid information loss by splitting unique sign pairs into individual histograms. We evaluate these variations with models trained on nine datasets spanning a range of drug target classes. Both 2DA_Sign and 3DA_Sign significantly increase model performance across all datasets when compared with traditional 2DA and 3DA. Lastly, we find that limiting 3DA_Sign to maximum atom pair distances of 6 Å instead of 12 Å further increases model performance, suggesting that conformational flexibility may hinder performance with longer 3DA descriptors. Consistent with this finding, limiting the number of bonds in 2DA_Sign from 11 to 5 fails to improve performance.

BCL::Conf: small molecule conformational sampling using a knowledge based rotamer library.

The interaction of a small molecule with a protein target depends on its ability to adopt a three-dimensional structure that is complementary. Therefore, complete and rapid prediction of the conformational space a small molecule can sample is critical for both structure- and ligand-based drug discovery algorithms such as small molecule docking or three-dimensional quantitative structure-activity relationships. Here we have derived a database of small molecule fragments frequently sampled in experimental structures within the Cambridge Structure Database and the Protein Data Bank. Likely conformations of these fragments are stored as 'rotamers' in analogy to amino acid side chain rotamer libraries used for rapid sampling of protein conformational space. Explicit fragments take into account correlations between multiple torsion bonds and effect of substituents on torsional profiles. A conformational ensemble for small molecules can then be generated by recombining fragment rotamers with a Monte Carlo search strategy. BCL::Conf was benchmarked against other conformer generator methods including Confgen, Moe, Omega and RDKit in its ability to recover experimentally determined protein bound conformations of small molecules, diversity of conformational ensembles, and sampling rate. BCL::Conf recovers at least one conformation with a root mean square deviation of 2 Å or better to the experimental structure for 99 % of the small molecules in the Vernalis benchmark dataset. The 'rotamer' approach will allow integration of BCL::Conf into respective computational biology programs such as Rosetta.Graphical abstract:Conformation sampling is carried out using explicit fragment conformations derived from crystallographic structure databases. Molecules from the database are decomposed into fragments and most likely conformations/rotamers are used to sample correspondng sub-structure of a molecule of interest.

Identification of specific ligand-receptor interactions that govern binding and cooperativity of diverse modulators to a common metabotropic glutamate receptor 5 allosteric site.

A common metabotropic glutamate receptor 5 (mGlu5) allosteric site is known to accommodate diverse chemotypes. However, the structural relationship between compounds from different scaffolds and mGlu5 is not well understood. In an effort to better understand the molecular determinants that govern allosteric modulator interactions with mGlu5, we employed a combination of site-directed mutagenesis and computational modeling. With few exceptions, six residues (P654, Y658, T780, W784, S808, and A809) were identified as key affinity determinants across all seven allosteric modulator scaffolds. To improve our interpretation of how diverse allosteric modulators occupy the common allosteric site, we sampled the wealth of mGlu5 structure-activity relationship (SAR) data available by docking 60 ligands (actives and inactives) representing seven chemical scaffolds into our mGlu5 comparative model. To spatially and chemically compare binding modes of ligands from diverse scaffolds, the ChargeRMSD measure was developed. We found a common binding mode for the modulators that placed the long axes of the ligands parallel to the transmembrane helices 3 and 7. W784 in TM6 not only was identified as a key NAM cooperativity determinant across multiple scaffolds, but also caused a NAM to PAM switch for two different scaffolds. Moreover, a single point mutation in TM5, G747V, altered the architecture of the common allosteric site such that 4-nitro-N-(1,3-diphenyl-1H-pyrazol-5-yl)benzamide (VU29) was noncompetitive with the common allosteric site. Our findings highlight the subtleties of allosteric modulator binding to mGlu5 and demonstrate the utility in incorporating SAR information to strengthen the interpretation and analyses of docking and mutational data.

Benchmarking ligand-based virtual High-Throughput Screening with the PubChem database.

With the rapidly increasing availability of High-Throughput Screening (HTS) data in the public domain, such as the PubChem database, methods for ligand-based computer-aided drug discovery (LB-CADD) have the potential to accelerate and reduce the cost of probe development and drug discovery efforts in academia. We assemble nine data sets from realistic HTS campaigns representing major families of drug target proteins for benchmarking LB-CADD methods. Each data set is public domain through PubChem and carefully collated through confirmation screens validating active compounds. These data sets provide the foundation for benchmarking a new cheminformatics framework BCL::ChemInfo, which is freely available for non-commercial use. Quantitative structure activity relationship (QSAR) models are built using Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), Decision Trees (DTs), and Kohonen networks (KNs). Problem-specific descriptor optimization protocols are assessed including Sequential Feature Forward Selection (SFFS) and various information content measures. Measures of predictive power and confidence are evaluated through cross-validation, and a consensus prediction scheme is tested that combines orthogonal machine learning algorithms into a single predictor. Enrichments ranging from 15 to 101 for a TPR cutoff of 25% are observed.

Calcium-activated nonspecific cation current and synaptic depression promote network-dependent burst oscillations.

Central pattern generators (CPGs) produce neural-motor rhythms that often depend on specialized cellular or synaptic properties such as pacemaker neurons or alternating phases of synaptic inhibition. Motivated by experimental evidence suggesting that activity in the mammalian respiratory CPG, the preBötzinger complex, does not require either of these components, we present and analyze a mathematical model demonstrating an unconventional mechanism of rhythm generation in which glutamatergic synapses and the short-term depression of excitatory transmission play key rhythmogenic roles. Recurrent synaptic excitation triggers postsynaptic Ca(2+)-activated nonspecific cation current (I(CAN)) to initiate a network-wide burst. Robust depolarization due to I(CAN) also causes voltage-dependent spike inactivation, which diminishes recurrent excitation and thus attenuates postsynaptic Ca(2+) accumulation. Consequently, activity-dependent outward currents-produced by Na/K ATPase pumps or other ionic mechanisms-can terminate the burst and cause a transient quiescent state in the network. The recovery of sporadic spiking activity rekindles excitatory interactions and initiates a new cycle. Because synaptic inputs gate postsynaptic burst-generating conductances, this rhythm-generating mechanism represents a new paradigm that can be dubbed a 'group pacemaker' in which the basic rhythmogenic unit encompasses a fully interdependent ensemble of synaptic and intrinsic components. This conceptual framework should be considered as an alternative to traditional models when analyzing CPGs for which mechanistic details have not yet been elucidated.