Generalization by symbolic abstraction in cascaded recurrent networks

Generalization performance in recurrent neural networks is enhanced by cascading several networks. By discretizing abstractions induced in one network, other networks can operate on a coarse symbolic level with increased performance on sparse and structural prediction tasks. The level of systematicity exhibited by the cascade of recurrent networks is assessed on the basis of three language domains. (C) 2004 Elsevier B.V. All rights reserved.

Interaction of the Hsp90 cochaperone cyclophilin 40 with Hsc70

The high-affinity ligand-binding form of unactivated steroid receptors exists as a multicomponent complex that includes heat shock protein (Hsp)90; one of the immunophilins cyclophilin 40 (CyP40), FKBP51, or FKBP52; and an additional p23 protein component. Assembly of this heterocomplex is mediated by Hsp70 in association with accessory chaperones Hsp40, Hip, and Hop. A conserved structural element incorporating a tetratricopeptide repeat (TPR) domain mediates the interaction of the immunophilins with Hsp90 by accommodating the C-terminal EEVD peptide of the chaperone through a network of electrostatic and hydrophobic interactions. TPR cochaperones recognize the EEVD structural motif common to both Hsp90 and Hsp70 through a highly conserved clamp domain. In the present study, we investigated in vitro the molecular interactions between CyP40 and FKBP52 and other stress-related components involved in steroid receptor assembly, namely Hsp70 and Hop. Using a binding protein-retention assay with CyP40 fused to glutathione S-transferase immobilized on glutathione-agarose, we have identified the constitutively expressed form of Hsp70, heat shock cognate (Hsc)70, as an additional target for CyP40. Deletion mapping studies showed the binding determinants to be similar to those for CyP40-Hsp90 interaction. Furthermore, a mutational analysis of CyP40 clamp domain residues confirmed the importance of this motif in CyP40-Hsc70 interaction. Additional residues thought to mediate binding specificity through hydrophobic interactions were also important for Hsc70 recognition. CyP40 was shown to have a preference for Hsp90 over Hsc70. Surprisingly, FKBP52 was unable to compete with CyP40 for Hsc70 binding, suggesting that FKBP52 discriminates between the TPR cochaperone-binding sites in Hsp90 and Hsp70. Hop, which contains multiple units of the TPR motif, was shown to be a direct competitor with CyP40 for Hsc70 binding. Similar to Hop, CyP40 was shown not to influence the adenosine triphosphatase activity of Hsc70. Our results suggest that CyP40 may have a modulating role in Hsc70 as well as Hsp90 cellular function.

Neuromuscular adaptation during skill acquisition on a two degree-of-freedom target-acquisition task: Isometric torque production

In this study we attempted to identify the principles that govern the changes in neural control that occur during repeated performance of a multiarticular coordination task. Eight participants produced isometric flexion/extension and pronation/supination torques at the radiohumeral joint, either in isolation (e.g., flexion) or in combination (e.g., flexion - supination), to acquire targets presented by a visual display. A cursor superimposed on the display provided feedback of the applied torques. During pre- and postpractice tests, the participants acquired targets in eight directions located either 3.6 cm (20% maximal voluntary contraction [MVC]) or 7.2 cm (40% MVC) from a neutral cursor position. On each of five consecutive days of practice the participants acquired targets located 5.4 cm (30% MVC) from the neutral position. EMG was recorded from eight muscles contributing to torque production about the radiohumeral joint during the pre- and posttests. Target-acquisition time decreased significantly with practice in most target directions and at both target torque levels. These performance improvements were primarily associated with increases in the peak rate of torque development after practice. At a muscular level, these changes were brought about by increases in the rates of recruitment of all agonist muscles. The spatiotemporal organization of muscle synergies was not significantly altered after practice. The observed adaptations appear to lead to performances that are generalizable to actions that require both greater and smaller joint torques than that practiced, and may be successfully recalled after a substantial period without practice. These results suggest that tasks in which performance is improved by increasing the rate of muscle activation, and thus the rate of joint torque development, may benefit in terms of the extent to which acquired levels of performance are maintained over time.

Neuromuscular adaptation during skill acquisition on a two degree-of-freedom target-acquisition task: Dynamic movement

In this experiment, we examined the extent to which the spatiotemporal reorganization of muscle synergies mediates skill acquisition on a two degree-of-freedom (df) target-acquisition task. Eight participants completed five practice sessions on consecutive days. During each session they practiced movements to eight target positions presented by a visual display. The movements required combinations of flexion/extension and pronation/supination of the elbow joint complex. During practice sessions, eight targets displaced 5.4 cm from the start position ( representing joint excursions of 54) were presented 16 times. During pre- and posttests, participants acquired the targets at two distances (3.6 cm [36 degrees] and 7.2 cm [72 degrees]). EMG data were recorded from eight muscles contributing to the movements during the pre- and posttests. Most targets were acquired more rapidly after the practice period. Performance improvements were, in most target directions, accompanied by increases in the smoothness of the movement trajectories. When target acquisition required movement in both dfs, there were also practice-related decreases in the extent to which the trajectories deviated from a direct path to the target. The contribution of monofunctional muscles ( those producing torque in a single df) increased with practice during movements in which they acted as agonists. The activity in bifunctional muscles ( those contributing torque in both dfs) remained at pretest levels in most movements. The results suggest that performance gains were mediated primarily by changes in the spatial organization of muscles synergies. These changes were expressed most prominently in terms of the magnitude of activation of the monofunctional muscles.

Tertiapin-Q blocks recombinant and native large conductance K+ channels in a use-dependent manner

Tertiapin, a short peptide from honey bee venom, has been reported to specifically block the inwardly rectifying K+ (Kir) channels, including G protein-coupled inwardly rectifying potassium channel (GIRK) 1 + GIRK4 heteromultimers and ROMK1 homomultimers. In the present study, the effects of a stable and functionally similar derivative of tertiapin, tertiapin-Q, were examined on recombinant human voltage-dependent Ca2+-activated large conductance K+ channel (BK or MaxiK; alpha-subunit or hSlo1 homomultimers) and mouse inwardly rectifying GIRK1 + GIRK2 (i.e., Kir3.1 and Kir3.2) heteromultimeric K+ channels expressed in Xenopus oocytes and in cultured newborn mouse dorsal root ganglion (DRG) neurons. In two-electrode voltage-clamped oocytes, tertiapin-Q (1-100 nM) inhibited BK-type K+ channels in a use- and concentration-dependent manner. We also confirmed the inhibition of recombinant GIRK1 + GIRK2 heteromultimers by tertiapin-Q, which had no effect on endogenous depolarization- and hyperpolarization-activated currents sensitive to extracellular divalent cations (Ca2+, Mg2+, Zn2+, and Ba2+) in defolliculated oocytes. In voltage-clamped DRG neurons, tertiapin-Q voltage- and use-dependently inhibited outwardly rectifying K+ currents, but Cs+-blocked hyperpolarization-activated inward currents including I-H were insensitive to tertiapin-Q, baclofen, barium, and zinc, suggesting absence of functional GIRK channels in the newborn. Under current-clamp conditions, tertiapin-Q blocked the action potential after hyperpolarization (AHP) and increased action potential duration in DRG neurons. Taken together, these results demonstrate that the blocking actions of tertiapin-Q are not specific to Kir channels and that the blockade of recombinant BK channels and native neuronal AHP currents is use-dependent. Inhibition of specific types of Kir and voltage-dependent Ca2+-activated K+ channels by tertiapin-Q at nanomolar range via different mechanisms may have implications in pain physiology and therapy.

A versatile synthetic approach to peptidyl privileged structures using a "safety-catch" linker

Peptidyl privileged structures have been widely used by many groups to discover biologically active molecules. In this context, privileged substructures are used as hydrophobic anchors, to which peptide functionality is appended to gain specificity. Utilization of this concept has led to the discovery of many different active compounds at a wide range of biological receptors. A synthetic approach to these compounds has been developed on a safety-catch linker that allows rapid preparation of large libraries of these molecules. Importantly, amide bond formation/cleavage through treatment with amines is the final step; it is a linker strategy that allows significant diversification to be easily incorporated, and it only requires the inclusion of an amide bond. In addition, chemistry has been developed that permits the urea moiety to be inserted at the N-terminus of the peptide, allowing the same set of amines (either privileged substructures or amino acid analogues) to be used at both the N- and C-termini of the molecule. To show the robustness of this approach, a small library of peptidyl privileged structures were synthesized, illustrating that large combinatorial libraries can be synthesized using these technologies.

Prediction of subcellular localization using sequence-biased recurrent networks

Motivation: Targeting peptides direct nascent proteins to their specific subcellular compartment. Knowledge of targeting signals enables informed drug design and reliable annotation of gene products. However, due to the low similarity of such sequences and the dynamical nature of the sorting process, the computational prediction of subcellular localization of proteins is challenging. Results: We contrast the use of feed forward models as employed by the popular TargetP/SignalP predictors with a sequence-biased recurrent network model. The models are evaluated in terms of performance at the residue level and at the sequence level, and demonstrate that recurrent networks improve the overall prediction performance. Compared to the original results reported for TargetP, an ensemble of the tested models increases the accuracy by 6 and 5% on non-plant and plant data, respectively.

Neuromuscular-skeletal constraints on the acquisition of skill in a discrete torque production task

The organisation of the human neuromuscular-skeletal system allows an extremely wide variety of actions to be performed, often with great dexterity. Adaptations associated with skill acquisition occur at all levels of the neuromuscular-skeletal system although all neural adaptations are inevitably constrained by the organisation of the actuating apparatus (muscles and bones). We quantified the extent to which skill acquisition in an isometric task set is influenced by the mechanical properties of the muscles used to produce the required actions. Initial performance was greatly dependent upon the specific combination of torques required in each variant of the experimental task. Five consecutive days of practice improved the performance to a similar degree across eight actions despite differences in the torques required about the elbow and forearm. The proportional improvement in performance was also similar when the actions were performed at either 20 or 40% of participants' maximum voluntary torque capacity. The skill acquired during practice was successfully extrapolated to variants of the task requiring more torque than that required during practice. We conclude that while the extent to which skill can be acquired in isometric actions is independent of the specific combination of joint torques required for target acquisition, the nature of the kinetic adaptations leading to the performance improvement in isometric actions is influenced by the neural and mechanical properties of the actuating muscles.

STAR: predicting recombination sites from amino acid sequence

Background: Designing novel proteins with site-directed recombination has enormous prospects. By locating effective recombination sites for swapping sequence parts, the probability that hybrid sequences have the desired properties is increased dramatically. The prohibitive requirements for applying current tools led us to investigate machine learning to assist in finding useful recombination sites from amino acid sequence alone. Results: We present STAR, Site Targeted Amino acid Recombination predictor, which produces a score indicating the structural disruption caused by recombination, for each position in an amino acid sequence. Example predictions contrasted with those of alternative tools, illustrate STAR'S utility to assist in determining useful recombination sites. Overall, the correlation coefficient between the output of the experimentally validated protein design algorithm SCHEMA and the prediction of STAR is very high (0.89). Conclusion: STAR allows the user to explore useful recombination sites in amino acid sequences with unknown structure and unknown evolutionary origin. The predictor service is available from http://pprowler.itee.uq.edu.au/star.

Detecting and sorting targeting peptides wtih neural networks and support vector machines

This paper presents a composite multi-layer classifier system for predicting the subcellular localization of proteins based on their amino acid sequence. The work is an extension of our previous predictor PProwler v1.1 which is itself built upon the series of predictors SignalP and TargetP. In this study we outline experiments conducted to improve the classifier design. The major improvement came from using Support Vector machines as a "smart gate" sorting the outputs of several different targeting peptide detection networks. Our final model (PProwler v1.2) gives MCC values of 0.873 for non-plant and 0.849 for plant proteins. The model improves upon the accuracy of our previous subcellular localization predictor (PProwler v1.1) by 2% for plant data (which represents 7.5% improvement upon TargetP).

Predicting structural disruption of proteins caused by crossover

We present a machine learning model that predicts a structural disruption score from a protein’s primary structure. SCHEMA was introduced by Frances Arnold and colleagues as a method for determining putative recombination sites of a protein on the basis of the full (PDB) description of its structure. The present method provides an alternative to SCHEMA that is able to determine the same score from sequence data only. Circumventing the need for resolving the full structure enables the exploration of yet unresolved and even hypothetical sequences for protein design efforts. Deriving the SCHEMA score from a primary structure is achieved using a two step approach: first predicting a secondary structure from the sequence and then predicting the SCHEMA score from the predicted secondary structure. The correlation coefficient for the prediction is 0.88 and indicates the feasibility of replacing SCHEMA with little loss of precision. ©2005 IEEE