7 More recently, the findings of a bioassay showed that the ZOL c

7 More recently, the findings of a bioassay showed that the ZOL concentrations found in the oral

cavity of patients under treatment with this drug ranged from 0.4 to 5 μM.17 Thereafter, some authors have demonstrated that this drug can be toxic to different cell types, such as osteoblasts, endothelial cells and fibroblasts.10, 11 and 16 This cytotoxic effect could be due to contact of high concentrations of bisphosphonates released from the mineralized tissues to the adjacent cells. A recent study5 evaluating the cytotoxicity of ZOL to pulp cells in vitro showed that this drug caused a significant decrease of the viability, proliferation RG7422 ic50 and TP production of these cells. These data were confirmed in the present study in which ZOL

concentrations (1 μM and 5 μM) simulating those found in the alveolar bone tissue of patients under treatment with this drug, 17 caused reduction of cell viability. In addition to GDC-0199 manufacturer the analysis of odontoblast-like cell viability, TP production and ALP activity, molecular biology experiments were also carried out in the present study, which indicated that Col-I and ALP expression can be inhibited in a dose-dependent by the action of ZOL. This inhibitory effect of ZOL could affect negatively the repair of the pulpo-dentin complex in vivo, as Col-I is the main component of reactionary dentin matrix, which is produced by odontoblasts ifenprodil that suffer aggressions 12 and 27 and ALP is directly involved in the mineralization of this newly formed dentin matrix. 28 and 29 The present study demonstrated that ZOL at concentration of 1 μM increase Col-I expression. Similar result was also reported in previous studies that revealed an increase in the expression of this gene in vitro within the first days after contact of the drug with the cells. 30 and 31 However, Col-I expression decreased over time,

suggesting that the inhibitory effect of ZOL was both dose- and time-dependent. 31 The results of ZOL cytotoxicity to the odontoblast-like MDPC-23 cells demonstrated by the in vitro cellular and molecular biology protocols used in the present study were confirmed in the analysis of cell morphology by SEM. The cells incubated in contact with both ZOL concentrations presented size reduction, probably due to cytoskeletal shrinkage. This cell response pattern in contact with low toxic agents has been extensively described in the literature, 19 and 21 which helps establishing the effects of the tested drug. This is because the decrease of cell viability indicated by the MTT assay might be due to a direct inhibitory effect of the drug on cell activity, which results in reversible morphological alterations, or to necrotic or apoptotic cell death, which represents an irreversible condition. In both situations, the MTT assay provides values that represent a smaller number of formazan crystals formed.

For validation of the method, we test the performance of our trai

For validation of the method, we test the performance of our training approach on a reference dataset of kinematic variables of human walking motion and compare

it against the existing TRBM model and the Conditional RBM (CRBM) as a benchmark (Taylor et al., 2007). As an application of our model, we train the TRBM using temporal autoencoding on natural movie sequences and find that the neural elements develop dynamic RFs that Birinapant purchase express smooth transitions, i.e. translations and rotations, of the static receptive field model. Our model neurons account for spatially and temporally sparse activities during stimulation with natural image sequences and we demonstrate this by simulation of neuronal spike train responses driven by the dynamic model responses. Our results propose how neural dynamic RFs may emerge naturally from smooth image sequences. We outline a novel method to learn temporal

and spatial structure from dynamic stimuli – in our case smooth image sequences – with artificial neural networks. The hidden units (neurons) of these generative models develop dynamic RFs that represent smooth temporal evolutions of static RF models that have been described previously for natural still images. When stimulated with natural movie sequences the model units are activated sparsely, both in space and time. A point process model translates the model’s unit activation isocitrate dehydrogenase targets into sparse neuronal spiking activity with few neurons being active at any given point in time and sparse single neuron firing patterns. We rely on the general model class of RBMs (see Section 4.1). The classic RBM is a two layer artificial neural network with a visible   and a hidden   layer used to learn representations of a dataset in an unsupervised fashion ( Fig. 1A). The units

(neurons) in the visible   and those in the hidden   layers are all-to-all connected Gefitinib via symmetric weights and there is no connectivity between neurons within the same layer. The input data, in our case natural images, activate the units of the visible   layer. This activity is then propagated to the hidden   layer where each neuron’s activity is determined by the input data and by the weights WW connecting the two layers. The weights define each hidden neuron’s filter properties or its RF, determining its preferred input. Whilst the RBM has been successfully used to model static data, it lacks in the ability to explicitly represent the temporal evolution of a continuous dataset. The CRBM (Fig. 1C) and TRBM (Fig. 1D) are both temporal extensions of the RBM model, allowing the hidden unit activations to be dependent on multiple samples of a sequential dataset. The models have a delay parameter which is used to determine how long the integration period on a continuous dataset is.

e we aim to identify all the different names in use for an enzym

e. we aim to identify all the different names in use for an enzyme and collect this information at one place: the BRENDA database (Chang et al., 2009 and Scheer et al., 2011). During the manual Dapagliflozin purchase annotation or the literature search the curators extract systematically all names and synonyms that are used for a specific enzyme except those that are totally meaningless (such as quantum for EC, or HAT for, or DDT for EC These are in later update rounds used as search terms for the identification of relevant literature. As a result

BRENDA is good source for enzyme synonyms storing about 82,000 different enzyme names for the around 5200 enzymes classified. This number clearly shows the dramatic problems: on average each EC class is recorded with 15

different names. This means that a literature search with any particular GSK1120212 clinical trial enzyme name on average finds only 1/15, i.e., less than 8% of the relevant literature. Only 20% out of the EC classes are listed with only the accepted name plus a systematic name. 10% out of the EC classes carry only one synonym and 40% are recorded with 2–5 synonyms. Looking at these enzymes it is a general observation that enzymes with a low number of synonyms very often possess a rather narrow substrate specificity or even are specific for a single substrate. Some have been identified in the secondary metabolism of a single plant and are absent from plants in taxonomically related species. 61 EC classes are stored with more than 100 different names, where 30 have more than 150 names (see Table 1). There are different reasons for the large number of different names. If we consider the protein kinases we find very high numbers of synonyms, each for an individual protein catalysing the phosphate transfer either to tyrosine, serine, threonine or histidine. Since the reaction which is the basis for classification is identical, the enzymes are assembled under just a few EC numbers but are named for the individual role they play in different organisms. In organism 1 Mannose-binding protein-associated serine protease they could, e.g., phosphorylate a specific protein at a specific position, in organism 2 the same enzyme could phosphorylate

a different protein. As long as the substrate specificity is not thoroughly analysed they are classified in the same EC-number. This could change in the future once it is proven that they have distinctly different substrate specificities. It is obvious from the table that especially for enzymes modifying proteins or other macromolecules many different names are in use. A different situation is found in the cellulase case, for example. The number of different substrates accepted here is very small, being mainly amorphous or crystalline cellulose. 220 different names are presently in use in the literature. In this case the cellulose breakdown is achieved by a combination/cooperation of a number of isoenzymes. For these isoenzymes different terms are in use in the different organisms.

Milkov (2004) conservatively estimated global methane hydrate sou

Milkov (2004) conservatively estimated global methane hydrate sources to be composed of ca. 1–5×1015 m3 in terms of methane. This amount of hydrated gas is approximately twice as much Ku0059436 as that of natural gas present in all hydrocarbon reservoirs (Sloan and Koh, 2007). Methane in these reservoirs is mostly of biogenic origin (Koh et al., 2011). Hence, studies on methanogens associated with methane hydrate reservoirs are important.

A methanogen was isolated from deep sub seafloor methane hydrate sediment from the Krishna Godavari Basin off the eastern coast of India, following enrichment in MS medium (Boone et al., 1989) with H2 and CO2 as a source of carbon and energy and subsequent isolation using the roll tube method (Hungate, 1950). This isolate (designated as click here MH98A) was identified as a putative novel species of the genus Methanoculleus on the basis of its mcrA gene and 16S rRNA gene sequence featuring similarities of 94% and 99% respectively with the closest phylogenetic relative, Methanoculleus marisnigri JR1 (GenBank Accession No. NC_009051.1; Anderson et al.,

2009). Similar enrichment and isolation of methanogens was performed using MS medium supplemented with alternate substrates such as formate, acetate, methylamine and methanol. However, all isolates showed a similar phylogenetic affiliation. Hence, strain MH98A was believed to be the dominant methanogen principally contributing to methane hydrate deposits in the Krishna Godavari basin. Considering the enormous volumes of methane hydrate deposits in the region and Methanoculleus sp. MH98A as a dominant methanogen, gaining insights into the genome organization of MH98A was of immense interest to understand the methanogenesis that almost entirely contributes to the

vast methane hydrate deposits. Characterization of the methanogenic metabolism of this organism is crucial to deduce the magnitude and the energy content of methane hydrate deposits. To our best knowledge, genome sequences Amisulpride of other methanogens associated with deep submarine methane hydrate deposits are not available so far. Further studies on these kinds of microorganisms to exploit their massive methanogenic potential could possibly revolutionize the energy industry. The genome of strain MH98A was sequenced using the Ion Torrent PGM sequencer (200-bp library) applying the 316™ sequencing chip according to the manufacturer’s instructions (Life Technologies, USA). De novo assembly was performed using version 4.0.5 of MIRA Assembler ( Chevreux et al., 1999) and generated 80 large contigs (> 8000 bp) and 226 smaller contigs (< 8000 bp) featuring a G + C content of 61.4%, an N50 value of 27533 bp, an N90 value of 4146 bp and a maximum contig size of 135,061 bp ( Table 1). All of 306 contigs were used for gene prediction and annotation by the RAST (Rapid Annotation using Subsystem Technology) system ( Aziz et al., 2008), with tRNAscan-SE-1.23 software ( Lowe and Eddy, 1997). RAST analysis revealed that, M.

Although rhLK8 significantly reduced tumor size in a limited peri

Although rhLK8 significantly reduced tumor size in a limited period of time (~ 4 weeks) by inducing apoptosis of tumor-associated endothelial cells, leading to the induction of apoptosis of nearby tumor cells nourished by the same vasculature, it did not affect tumor cell proliferation. These findings suggest that the cytostatic nature of angiogenesis inhibitors, including rhLK8, may limit their ability to control the growth of cancer cells,

and combination therapy with chemotherapeutic agents may be necessary to enhance their therapeutic efficacy and to prolong BMS-354825 price the median survival of patients with ovarian cancer. In this study, we found that antiangiogenic and antitumor efficacy was dramatically improved in mice treated with the combination of paclitaxel and rhLK8.

Our results are in agreement with an increasing body of work demonstrating that the combination of angiogenesis inhibitors with chemotherapeutic drugs significantly improves treatment outcomes compared to single agent therapy. Tumor blood vessels are irregular, dilated, tortuous, and leaky, which leads to elevated tumor interstitial fluid pressure and thus inefficient delivery of chemotherapeutic agents [36]. Antiangiogenic therapy may induce the transient normalization of tumor vasculature, which enhances the delivery Rapamycin price of chemotherapeutic agents such as paclitaxel enough by decreasing interstitial pressure, leading to an increase in therapeutic efficacy [37]. In addition, rhLK8 may attenuate the survival pathway of tumor-associated endothelial cells, which makes proliferating tumor-associated endothelial cells more sensitive

to anticycling drug, paclitaxel. The improved therapeutic outcomes induced by combination therapy with rhLK8 appear not to be limited to taxane-based chemotherapy. Our preliminary data showed that combination therapy of rhLK8 with gemcitabine or irinotecan (or 5-fluorouracil) improved treatment outcomes than the corresponding treatment as a single agent in the human pancreatic and colon carcinoma animal models, respectively (Kim JS et al., unpublished data). In this context, combination of rhLK8 with other chemotherapeutic agents such as carboplatin or cisplatin, which have been regarded as the primary treatment option of advanced ovarian cancer together with paclitaxel, was also expected to show improved treatment outcomes, although appropriate preclinical and/or clinical evaluation of the combination therapy will be critically required. Paclitaxel significantly reduced the volume of ascites in SKOV3ip1 tumor–bearing mice but rhLK8 alone did not. However, the effect of rhLK8 on decreasing MVD was more significant than that of paclitaxel.

The selected list of publications was also analyzed according to

The selected list of publications was also analyzed according to the distribution of the assay information and checked for different formats in which these data are represented in the Cyclopamine molecular weight publications. In more than 90% of the papers the assay conditions are described in free text, mainly within the Material and Methods section. But about 50% of the publications also represent assay conditions in the legends of tables

or figures. And a similar amount includes compound concentrations as part of the assay conditions within figures so that concentrations have to be extracted from graph axes. In some cases there are conflicts between information written in the free text of the Material and Methods section and assay conditions represented

in the legends of tables or figures. Within the set of analyzed articles we found two papers containing such conflicts. To solve these problems curators try to contact the authors where possible. Often the Material and Methods section contains a general description of the assay method and the legends contain more detailed or modified information about the experimental conditions for the measurement of the parameters displayed in the table or figure. One of our main interests in the paper analysis was the question how exact the entities (e.g. proteins, Panobinostat cost enzymes) can be identified within an article. The outcome was very surprising. We know that some older papers have incomplete data due to the lack of the state of the art at the time. For example, a definite identification of isozymes is often missing in old publications because it was simply not known at that time point that different isozymes exist. In the 1980s three main data resources were available and evolved as standard repositories for nucleotides and proteins: the Protein Data Bank (PDB) (Berman, 2008), SwissProt/UniProtKB (The UniProt Consortium, 2011) and the International Nucleotide Sequence Database Collection (INSDC) comprised of the three databases

DDBJ/EMBL/GenBank (Nakamura et al., 2013). Based on the availability of Y-27632 2HCl such standard protein and gene databases authors now have the possibility to exactly assign proteins to specific known isozymes by using database accession numbers. Additionally, starting in the 1990s, online repositories for ontologies and controlled vocabularies were developed to establish a universal standard terminology in biology e.g. Gene Ontology (The Gene Ontology Consortium, 2000) or NCBI organism taxonomy. A defined vocabulary is important to avoid misinterpretations and helps to exchange data between resources correctly. Ontologies and hierarchical classifications structure the data of a specific domain, describe the objects and define relationships between these objects. The usage of unique identifiers given by ontologies, controlled vocabularies and databases is essential for a definite data assignment.

Half of the patients had T2 and half had Gleason 7 prostate cance

Half of the patients had T2 and half had Gleason 7 prostate cancer. They administered HDR in a single implant over 2 days in three fractions; four different dose schedules were evaluated (10, 10.5, 11, or 11.5 Gy). The 3- and 5-year check details biochemical control rates (nadir + 2) were 88% and 85%. There were no differences in toxicity between doses. Acute rectal toxicity was nearly all Grade 1 and acute Grade 3 urinary toxicity occurred in only 1 patient. Chronic Grade 3

urinary toxicity was <10% and no Grade 4 toxicities were recorded. The group from Offenbach Germany, lead by Zamboglou and Baltas, obtained excellent results in 718 patients using intraoperative TRUS treatment planning. The dose and fractionation schedule evolved over time (51). Protocol A (9.5 Gy × 4 in one implant), protocol B (9.5 Gy × 4

in two implants), and finally the current protocol C (11.5 Gy × 3 in three implants). The authors progressively included higher risk group cases so that for protocol C 57% of cases were intermediate- or high-risk compared with 27% in protocol A and 44% in protocol B. The median followup by protocol was 7.7 years for 141 patients (protocol A), 4.9 years for 351 patients (protocol B), and 2.1 years selleck chemical for 226 patients (protocol C). The 3-year biochemical control for all patients was 95% and distant metastasis–free survival was 98%. The 5-year results were available for protocols A and B (9.5 Gy × 4). Biochemical control was 97% and 94%. There were no significant differences correlated with T score, PSA, Gleason mafosfamide score, or risk group. Late Grade 3 GU and GI toxicities were 3.5% and 1.6%. Urinary strictures that required urethrotomy (Grade 3 GU toxicity) occurred in 1.8% and 2 patients required urinary diversion to manage urinary incontinence (Grade 4 GU toxicity). Although the followup is significantly less in protocol C, there were no apparent differences in tumor control or morbidity between

the three protocols. Ghilezan et al. (52) reported on an ultra-hypofractionated HDR monotherapy trial for low- and intermediate-risk prostate cancer that accrued 100 patients. The total dose was 24 Gy for the first 50 patients (one implant, two fractions, and 6 h interfraction interval) and 27 Gy in the next 50 patients. The median followup was 17 months. There were no differences in acute or chronic toxicities between the two doses. The maximum chronic GU and GI toxicities Grade 2 or higher were ≤5% with the exception of urinary frequency/urgency, which was 16%. These symptoms resolved by 6 months in most cases (0% for the 24 Gy and 4.8% for the 27 Gy). The program was changed to two implants 2–3 weeks apart to increase the time for normal tissue repair and to shorten the time of the procedure per day by removing the same day waiting between fractions.

5) These data suggest that the chemistry of each of the flow reg

5). These data suggest that the chemistry of each of the flow regimes is controlled

by different factors and/or combinations of factors. One plausible explanation for the differences in stormflow and baseflow water chemistry is the chemical variation imparted by differences in river water pH between the two events. The samples collected along the length of the river after Tropical Storm Irene had a mean pH value (5.54 ± 0.32), within analytical error of natural rainfall. Those collected during baseflow conditions are near neutral (6.86 ± 0.33). Both sampling events show relatively little chemical variation along the length of the river (Fig. 3 and Fig. 4), however, the slightly enhanced concentration of the relative insoluble elements, like Al, Fe, and the REEs during the stormflow sampling is see more attributed to this difference in pH. During both sampling events (stormflow r2 = 0.65; baseflow r2 = 0.70) pH increased slightly downriver ( Table 2 and Fig. 3) while specific

conductance fell during stormflow (r2 = −0.58) but rose during baseflow (0.38). Another factor click here which could drive the chemical differences between the two sampling events is the proportion of river water derived by overland versus groundwater flow. The water entering the river via runoff and overland flow after a heavy rainfall would follow shallow flow paths, have relatively little time for buffering and interaction with geologic materials, while discharge volumes would be many times those

occurring during baseflow, (∼14× in this comparison). In addition, in the Adirondack region, particularly the western portions, decades of acidic precipitation have leached the soil and sediment of soluble elements. Thus geological materials encountered by runoff and along shallow flow paths, have lost of much of their calcium, magnesium, and capacity to Oxalosuccinic acid buffer acidity (Jenkins et al., 2007, Lawrence, 2002, Lawrence et al., 2004, Lawrence et al., 2007 and Lawrence et al., 2008). During baseflow conditions water in a river system generally has longer and deeper flow paths, and more time to interact with geologic materials; some of which may be much less weathered than those at, or near, the surface. Baseflow should be better buffered and contain more of the elements with enhanced solubility at near neutral pH values, and approximate the composition of groundwater (Soulsby et al., 2003). The higher pH would also serve to limit the concentrations of most metals which have greater solubility in more acidic waters. Greater concentrations of anions (e.g. OH, CO3, and SO4) and higher pH would cause precipitation of insoluble phases containing metals such as Al, Fe, and the REEs. Carbonate dominates the anion population in both sampling events; however, the average concentrations during baseflow are almost twice those of stormflow conditions (12.35 vs. 6.99 mg/L), indicating more extensive interaction with carbonate-bearing geologic materials (Fig. 4).

LC was superior for extremity lesions compared with trunk tumors

LC was superior for extremity lesions compared with trunk tumors and HDR and EBRT compared with BT alone (odds ratio = 0.21; 95% confidence interval: 0.026, 0.651, p = 0.013). LC was also improved with doses greater than 65 Gy. A Japanese group reported their experience of HDR and EBRT. Their inclusion criteria were (1) high tumor grade, (2) low-grade

tumor ≥10 cm, (3) recurrent tumor, (4) tumor abutting or invading critical structures, and (5) positive margins. They prescribed 2–3 Gy/fraction × 6, BID combined with EBRT (36–60 Omipalisib Gy). After a median followup of 31 months, there was no local failure within the radiation field (25). San Miguel et al. (23) combined 45 Gy of EBRT with 16 or 24 Gy HDR BT depending on the margin status. LC at 9 years was reported as 77.4%. Positive margins had a 4.4-fold risk of local failure compared with close or negative margin (p = 0.036). They

report 30% Grade 3–4 toxic events, with the majority related to wound healing. Despite this relatively high rate of toxicity, the reoperation rate was comparable to other series at 10%. Lower limb location and volume of the 150% isodose (TV150 >27 mL) combined predicted for Grade 3 complications (p = 0.003). There is no randomized comparison of HDR and LDR BT. Pohar et al. (27), however, published a historical control comparison in 37 patients treated between 1995 and 2004. Twenty-seven patients had LDR and 17 patients HDR (since 2001). The mean EBRT dose check details was approximately 50 Gy. The LDR dose was 15 Gy prescribed at 6-mm depth (0.42 Gy/h) based on the Paris system of loading. The mean HDR dose was 13 Gy (10.2–18 Gy) over three to four fractions BID. They noted an increase in toxicity in patients receiving >15 Gy HDR and adopted a standard HDR dose of 4.5 Gy × 3 (13.5 Gy). LC was 90% with LDR and 94% for HDR. There was a trend of decreased

MycoClean Mycoplasma Removal Kit occurrence of severe complications (Grade 3–4) in the HDR group (30% LDR vs. 6% HDR p = 0.06). Laskar et al. (44) retrospectively reviewed their pediatric data for patients who underwent WLE with BT with or without EBRT. Both LDR and HDR were in their cohort. Of 50 patients, 30 had BT alone (LDR or HDR). They concluded that LC related to size of tumor and grade (better control for tumors <5 cm and low-grade tumors). LC for BT and EBRT was comparable to BT alone (78% vs. 84%, p = 0.89), and there was no difference in LC between LDR and HDR either as monotherapy or in combination with EBRT (77% vs. 92%, p = 0.32; 67% vs. 100%, p = 0.17). We concluded, therefore, that HDR is also a valid approach to source loading for STS. The radiobiology of large fraction sizes and the potential for creative combinations of HDR BT with systemic therapy is yet to be explored. HDR has some functional and radiation safety advantages for pediatric patients. There are a limited number of reports on the use of PDR BT in STS [28], [51] and [52].

Curves in Fig  2 show the behavior of the most thermal resistant

Curves in Fig. 2 show the behavior of the most thermal resistant between the curves from duplicate trials for each concentration. Table 3 summarizes the mean value and ABT-199 solubility dmso standard deviation of fitted parameter values, such β and α, and the t6D, at 100 °C and different EO concentrations (stage I). For the thermochemical resistance at 300 and 350 μg/g, the mean value of t6D was the same, these concentrations reduced the t6D in around 1.0 min from the thermal resistance without EO. The concentration of 400 μg/g resulted in a reduction of approximately 1.4 min and the concentration of 500 μg/g in 1.9 min in the t6D from the thermal resistance without EO. However, the

concentration of 400 μg/g was chosen to continue the experiment with different

temperatures since the organoleptic impact in a food product can be lower than at 500 μg/g. Subsequently, the thermochemical resistances were carried out with the fixed EO concentration of 400 μg/g and different temperatures. For the thermochemical resistance at 400 μg/g, the parameter mean values of β and α, and the mean value of t6D, with their respective standard deviation, are shown in Table 3 (stage II). As can been seen in Table 3, the values of parameter α for the thermochemical resistance at 400 μg/g of oregano EO do not depend on temperature since these values did not differ significantly Epigenetic pathway inhibitor with increasing temperature. Therefore, the Weibull model with a fixed α was fitted to the thermochemical experimental data. Some studies had already worked with the Weibull model with a fixed α ( Periago et al., 2004 and van Boekel, 2002) new achieving good results. The mean value of α for the thermochemical resistance with 400 μg/g of EO (stage II), equal to 2.65, was used to recalculate β and t6D. Fig. 3 exhibits the behavior of the most thermal resistant between the curves

from duplicate trials for each concentration generated through the Weibull model with parameter α fixed (2.65) with 400 μg/g of EO. The new mean values for parameter β and t6D, with their respective standard deviation, with constant α (2.65) and EO concentration (400 μg/g) are shown in Table 3 (stage III). Fig. 4 shows the dependence on temperature of the parameter β and the t6D for the Weibull model with fixed and varying α at 400 μg/g of oregano EO. Through Fig. 4, it can be observed that modeling with a fixed α did not significantly vary the values of β and t6D, similar to in the secondary model. Equations (5) and (6) show the secondary model for the temperature dependence of β and t6D with a fixed α, respectively. And Equations (7) and (8) present the secondary model for the temperature dependence of β and t6D with a varying α, respectively. The exponential equation (Equation (2)) showed a good fit to β and t6D, as can be seen in Fig. 4 and also through the R2 values. equation(5) β(T)=4.109exp(−0.21·T)R2=0.97 equation(6) t6D(T)=6·1010exp(−0.24·T)R2=0.97 equation(7) β(T)=2·109exp(−0.21·T)R2=0.