We describe a methodology for rapid experimentation in statistical machine translation which we use to add a large number of features to a baseline system exploiting features from a wide range of levels of syntactic representation. Feature values were combined in a log-linear model to select the highest scoring candidate translation from an n-best list. Feature weights were optimized directly against the BLEU evaluation metric on held-out data. We present results for a small selection of features at each level of syntactic representation.

Transcription is a stepwise process that involves many specialized proteins and protein complexes, all of which must work together to express a given gene in a spatially and temporally regulated manner. An integral step in this regulatory process is carried out by large, multisubunit co-activator complexes, which have diverse roles in transcriptional control. Their diversity and large size allows for many potential regulatory inputs, but how is the versatility and specificity of these co-activator complexes determined?

MHR3 is an ecdysone-inducible transcription factor whose expression in both Manduca sexta epidermis and the Manduca GV1 cell line is induced by 20-hydroxyecdysone (20E) in vitro. There are four putative ecdysone response elements (EcRE) in the 2.6-kb flanking region of the MHR3 promoter. The most proximal, EcRE1, is necessary for activation of the promoter by 20E in the GV1 cells because the mutation of EcRE1 caused the loss of responsiveness to 20E. Previous studies showed that EcR-B1/USP-1 bound only to EcRE1 and high levels of this complex increased the 20E-induced activation, whereas the presence of high USP-2 prevented this increased activation. When we expressed EcR-A alone or in combination with USP-1 under the control of Autographa californica baculovirus promoter (pIE1hr), the activation of the 2.6-kb promoter by 20E was reduced by about 50%. Moreover, when EcR-A was expressed together with both EcR-B1 and USP-1, it reduced the normal activation caused by EcR-B1 and USP-1 by 50%. Gel mobility shift assays showed no binding of EcR-A/USP-1 to EcRE1. The presence of EcR-A, however, reduced the binding of EcR-B1/USP-1 by about 50%. These findings suggest that EcR-A competes with EcR-B1 for binding of USP-1, leading to a decline in activity of the promoter. In addition, E75A, another ecdysone-induced transcription factor, and MHR3 itself suppressed MHR3 promoter activity by binding to the monomeric response element (MRE2). Therefore, MHR3 can be down-regulated both by itself and by E75A.

Sum frequency vibrational spectroscopy was used to study adsorption of leucine molecules at air-water interface from solutions with different concentrations and pH values. The surface density and the orientation of the isopropyl head group of the adsorbed leucine molecules could be deduced from the measurements. It was found that the orientation depends on the surface density, but only weakly on bulk pH value at the saturated surface density. The vibrational spectra of the interfacial water molecules appeared to be strongly affected by the charge state of the adsorbed leucine molecules. Enhancement and inversion of polar orientation of interfacial water molecules by surface charges or field controllable by the bulk pH value were observed.

Olfactory receptor neurons (ORNs) convey sensory information directly to the CNS via conventional glutamatergic synaptic contacts in olfactory bulb glomeruli. To better understand the process by which information contained in the odorant-evoked firing of ORNs is transmitted to the brain, we examined the properties of glutamate release from olfactory nerve (ON) terminals in slices of the rat olfactory bulb. We show that marked paired pulse depression is the same in simultaneously recorded periglomerular and tufted neurons, and that this form of short-term plasticity is attributable to a reduction of glutamate release from ON terminals. We used the progressive blockade of NMDA receptor (NMDAR) EPSCs by MK-801 [(5R,10S)-(+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5-10-imine hydrogen maleate] and stationary fluctuation analysis of AMPA receptor (AMPAR) EPSCs to determine the probability of release (P(r)) of ON terminals; both approaches indicated that P(r) is unusually high (>/=0.8). The low-affinity glutamate receptor antagonists gamma-d-glutamylglycine and l-amino-5-phosphonovaleric acid blocked ON-evoked AMPAR- and NMDAR-mediated EPSCs, respectively, to the same extent under conditions of low and high P(r), suggesting that multivesicular release is not a feature of ON terminals. Although release from most synapses exhibits a highly nonlinear dependence on extracellular Ca(2+), we find that the relationship between glutamate release and extracellular Ca(2+) at ON terminals is nearly linear. Our results suggest that ON terminals have specialized features that may contribute to the reliable transmission of sensory information from nose to brain.

Examination of the genetic differences between aphids that can transmit citrus tristeza virus, CTV, and those which cannot, may lead to a greater understanding of the virus-aphid interactions necessitating virus acquisition and transmission. Since a cDNA library had been completed the previous year for the brown citrus aphid, a vector of CTV, a second aphid cDNA library was made to a non-CTV aphid vector, the pea aphid, Acyrthosiphon pisum. Comparisons between these two genetic datasets will provide a better understanding of the dynamics of aphid feeding, digestion, development, and may elucidate elements related to virus interactions that were previously unknown. Identification of the numerous proteins actively involved in feeding and digestion from aphids will provide specific targets for the development of new methods of control aimed at disrupting aphid feeding and ultimately reducing the acquisition and transmission of plant viruses which cause disease.

The frontal and parietal eye fields serve as functional landmarks of the primate brain, although their correspondences between humans and macaque monkeys remain unclear. We conducted fMRI at 4.7 T in monkeys performing visually-guided saccade tasks and compared brain activations with those in humans using identical paradigms. Among multiple parietal activations, the dorsal lateral intraparietal area in monkeys and an area in the posterior superior parietal lobule in humans exhibited the highest selectivity to saccade directions. In the frontal cortex, the selectivity was highest at the junction of the precentral and superior frontal sulci in humans and in the frontal eye field (FEF) in monkeys. BOLD activation peaks were also found in premotor areas (BA6) in monkeys, which suggests that the apparent discrepancy in location between putative human FEF (BA6, suggested by imaging studies) and monkey FEF (BA8, identified by microstimulation studies) partly arose from methodological differences.

Insect molting is triggered by ecdysteroids, which are produced in the prothoracic glands (PG). The broad (br) gene is one of the ’early genes’ directly regulated by ecdysteroids. Ectopic expression of the BR-Z3 isoform in early second instar Drosophila larvae (L2) before the rise of the ecdysteroid titer prevented molting to the third instar, but the larvae subsequently formed L2 prepupae after prolonged feeding. When these larvae were fed on diet containing 20-hydroxyecdysone (20E), they formed pharate third instar larvae. The critical weight for normal L3 pupariation of w(1118) larvae was found to be 0.8 mg and that for L2 pupariation was 0.45 mg. We also defined a threshold weight for metamorphosis of 0.3 mg, above which L2 larvae will metamorphose when provided with 20E. BR-Z3 apparently works through the PG cells of the ring gland but not the putative neurosecretory cells that drive ecdysone secretion, because ectopic expression of BR-Z3 specifically in the ring gland caused 53% of the larvae to become permanent first instar larvae. Driving other BR isoforms in the ring gland prevented larval molting or pupariation to varying degrees. These molting defects were rescued by feeding 20E. Overexpression of each of the BR isoforms caused degeneration of the PG cells but on different time courses, indicating that BR is a signal for the degeneration of the PG cells that normally occurs during the pupal-adult transition.

Memory for object-location was investigated by testing subjects with small unilateral thermolesions to the medial temporal lobe using small-scale 2D (Abstract) or large-scale 3D (Real) recall conditions. Four patients with lesions of the left hippocampus (LH), 10 patients with damage to the right hippocampus (RH) and 9 matched normal controls (NC) were tested. Six task levels were presented in a pseudorandom order. During each level, subjects viewed one to six different objects on the floor of a circular curtained arena 2.90 m in diameter for 10 s. Recall was tested by marking the locations of objects on a map of the arena (Abstract recall) and then by replacing the objects in the arena (Real recall). Two component errors were studied by calculating the Location Error (LE), independent of the object identity and the configuration error by finding the best match to the presented configuration. The RH group was impaired relative to the NC for nearly all combinations of recall and error types. An impairment was observed in this group even for one object and it deepened sharply with an increasing object number. Damage to the right perirhinal or parahippocampal cortices did not add to the impairment. Deficits in the LH group were also observed, but less consistently. The data indicate that spatial memory is strongly but not exclusively lateralised to the right medial temporal lobe.