{"component": "clause", "props": {"groups": [{"snippet_links": [{"key": "other-conditions", "type": "definition", "offset": [206, 222]}, {"key": "field-of", "type": "definition", "offset": [292, 300]}, {"key": "effect-on-the", "type": "clause", "offset": [325, 338]}, {"key": "the-total", "type": "clause", "offset": [464, 473]}, {"key": "effect-of", "type": "clause", "offset": [474, 483]}, {"key": "reduction-in", "type": "definition", "offset": [515, 527]}, {"key": "the-field", "type": "definition", "offset": [688, 697]}, {"key": "the-yield", "type": "clause", "offset": [724, 733]}, {"key": "the-final", "type": "clause", "offset": [749, 758]}, {"key": "i-p", "type": "clause", "offset": [926, 929]}, {"key": "figure-7", "type": "clause", "offset": [957, 965]}, {"key": "changes-in", "type": "definition", "offset": [1489, 1499]}, {"key": "provide-a", "type": "definition", "offset": [1523, 1532]}, {"key": "caused-by", "type": "clause", "offset": [1579, 1588]}, {"key": "in-time", "type": "clause", "offset": [2255, 2262]}, {"key": "formation-of", "type": "clause", "offset": [2727, 2739]}, {"key": "in-the-case", "type": "clause", "offset": [3400, 3411]}, {"key": "the-exchange", "type": "clause", "offset": [3443, 3455]}, {"key": "for-example", "type": "clause", "offset": [4305, 4316]}], "samples": [{"hash": "aMJET9RloEz", "uri": "/contracts/aMJET9RloEz#figure-6", "label": "Doctoral Thesis", "score": 19.0, "published": true}], "size": 1, "snippet": "Relative 1O2 yield in Q-depleted reaction centers from Rb. sphaeroides R26 RCs as a function of the applied magnetic field. The inset shows the same measurements made over a wider range of magnetic fields. Other conditions as in Fig 5. obtained, shown in Figure 6. It appears that a magnetic field of a few mT has a profound effect on the 1O2 yield in R26 RCs: a 50 % reduction was observed for fields of 20-100 mT and even a 10 % reduction was observed for 1 mT. The total effect of 50% corresponds closely to the reduction in 3P yield observed at similar magnetic field strengths (\u2587\u2587\u2587\u2587\u2587\u2587\u2587 et al., 1985), and can thus be ascribed to the hyperfine mechanism outlined in the introduction. The field strength needed to change the yield by one half of the final change, B1/2, is 3.9 \u00b1 0.5 mT, which is slightly less than the B1/2-values found for the yield of 3P (4.2 mT and 5.7 mT) (\u2587\u2587\u2587\u2587\u2587\u2587\u2587 et al., 1985; \u2587\u2587\u2587\u2587\u2587 et al., 1986). B I P 0.4 Absorption 0.3 0.2 0.1 Figure 7 shows the absorption spectra of RCs in oxygen-saturated buffer before and after excitation with 12,000 laser flashes (\u03bbexc= 532 nm) with and without magnetic field. The illumination has irreversibly attenuated the absorption bands at 760, 800 and 860 nm belonging to I, the accessory bacteriochlorophyll B, and P respectively, and a slight absorption increase around 680 nm has occurred. These changes are indicative of photobleaching, accompanied by disruption of the interactions between the chromophores and possibly by changes in the RC structure. They provide a measure of the extent of the photodegradation caused by 1O2 (\u2587\u2587\u2587\u2587\u2587\u2587\u2587 et al., 2001). The bleaching of the 800 nm band is about 45% smaller in a field of 15 mT than it is in zero-field. This finding corroborates the measurements shown in Fig. 6 and demonstrates directly that a relatively modest magnetic field affords substantial protection for the RC against 1O2-induced damage. Wild-type reaction centers were much more stable under these light conditions and the total photobleaching was only 15 % and no magnetic field dependence was found (data not shown). We have demonstrated that in Q-depleted RCs from Rb. sphaeroides R26 the singlet oxygen yield after laser excitation is dependent on the applied magnetic field. In time-resolved measurements a 50 % decrease of singlet oxygen yield was measured and consequently a reduced photodegradation was observed in steady-state experiments. Thus the magnetic field protects this protein from photodegradation. Just a few milli-Tesla has a profound effect on the singlet oxygen yield while lifetimes remain constant, in the range of 40-46 \u00b5s. For an applied magnetic field to have a damaging effect in this context, it would need to promote the formation of 1O2. There are two ways in which a weak magnetic field (< 1 mT) could cause such an increase. First, there is the \u201cLow Field Effect\u201d which has an opposite phase to the effects reported here and occurs for fields smaller than the average hyperfine interactions in the radical pair (\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587 & \u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587, 1996; \u2587\u2587\u2587\u2587\u2587\u2587 et al., 1998). Second, and similar in appearance to the LFE, is the \u201c2J resonance\u201d that arises from energy-level crossings at field strengths that match the radical pair\u2019s exchange interaction (\u2587\u2587\u2587\u2587\u2587\u2587 & \u2587\u2587\u2587\u2587\u2587\u2587-\u2587\u2587\u2587\u2587\u2587\u2587\u2587, 1983). That neither of these machenisms operates efficiently here is due to the short lifetime of the radical pair and, in the case of the LFE, to the presence of the exchange and dipolar interactions between the two electron spins. A 2J resonance in the yield of 3P has been found for Rb. sphaeroides (\u2587\u2587\u2587\u2587\u2587\u2587 et al., 1987) but only at temperatures below 0 \u00b0C. In very strong magnetic fields (>5 T), the triplet yield becomes larger than in the absence of an applied field as a result of the difference in Zeeman interactions of the two radicals (\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587 & \u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587, 1996; \u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587 et al., 1988). The size of this effect and the field at which it occurs are determined by the difference in the two g-values, which is quite small for P+ and I\u2212 but can be much higher for other radical pairs, such that relatively modest fields could cause the photosensitised 1O2 yield to rise above that in zero field. These effects are not necessarily restricted to reaction centers or indeed to 1O2 produced by photosensitisation. For example, 1O2 is formed during lipid peroxidation by the self reaction of peroxyl radicals (\u2587\u2587\u2587 \u2587\u2587\u2587\u2587\u2587\u2587\u2587 mechanism): a process which could conceivably show RPM effects (\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587 et al., 2003).", "hash": "d52b7058550dc485eb7c9f828b1e7ece", "id": 1}, {"snippet_links": [{"key": "pass-through", "type": "definition", "offset": [57, 69]}, {"key": "the-site", "type": "clause", "offset": [70, 78]}, {"key": "development-site", "type": "definition", "offset": [186, 202]}, {"key": "in-general", "type": "clause", "offset": [219, 229]}], "samples": [{"hash": "958pBC7bq8c", "uri": "/contracts/958pBC7bq8c#figure-6", "label": "Consultancy Agreement", "score": 25.9322376251, "published": true}], "size": 1, "snippet": "1 shows the prevailing winds from ESE direction smoothly pass through the site and along Fanling highway. However, for the prevailing E wind, no potential wind corridors penetrating the development site can be captured in general.", "hash": "118e30cd57c61d0f80903bad5b9648b4", "id": 2}, {"snippet_links": [{"key": "before-starting", "type": "clause", "offset": [146, 161]}, {"key": "the-experiment", "type": "clause", "offset": [162, 176]}, {"key": "at-t", "type": "clause", "offset": [178, 182]}, {"key": "the-source", "type": "clause", "offset": [188, 198]}, {"key": "over-time", "type": "clause", "offset": [384, 393]}, {"key": "order-of", "type": "clause", "offset": [408, 416]}, {"key": "the-charges", "type": "clause", "offset": [716, 727]}], "samples": [{"hash": "bR47k9C1z3L", "uri": "/contracts/bR47k9C1z3L#figure-6", "label": "Licence Agreement Concerning Inclusion of Doctoral Thesis in the Institutional Repository of the University of Leiden", "score": 17.0, "published": true}], "size": 1, "snippet": "4: Single-molecule lines. The laser is scanned over 10 GHz. Vg is kept at -40 V during the whole experiment. Vsd was set at 50 V during 2 minutes before starting the experiment. At t = 0, the source-drain voltage was switched off. After 1 hour, we set it to 50 V and switched it off again after 30 minutes. The first regime exhibits molecules drifting towards a new spectral position over time scales of the order of a few hours (at least). These very slow drifts cover a range exceeding 10 GHz. In contrast with this, some molecules do not show any shift of their absorption frequencies. Similar effects had been reported in a former single-molecule study on silicon carbide (SiC) [21, 22]. However, in this study, the charges were not present in the matrix (which was a Shpol\u2019skii n-hexadecane matrix) but onto the surface of SiC, a large", "hash": "c040d1d1ea12e1e4c31fe02979b6f96e", "id": 3}, {"snippet_links": [{"key": "an-example", "type": "clause", "offset": [0, 10]}, {"key": "figure-5", "type": "definition", "offset": [66, 74]}], "samples": [{"hash": "2FRmd1iNVLh", "uri": "/contracts/2FRmd1iNVLh#figure-6", "label": "Textured Agreements", "score": 19.1697158813, "published": true}], "size": 1, "snippet": "An example of a vignette used to similar effect as the factoid in Figure 5.", "hash": "4ae6425c36173825b94f66510e0bfce8", "id": 4}, {"snippet_links": [{"key": "pass-through", "type": "definition", "offset": [56, 68]}, {"key": "the-site", "type": "clause", "offset": [69, 77]}], "samples": [{"hash": "958pBC7bq8c", "uri": "/contracts/958pBC7bq8c#figure-6", "label": "Consultancy Agreement", "score": 25.9322376251, "published": true}], "size": 1, "snippet": "2 shows the prevailing winds from SE and SSE directions pass through the site to ventilate the downstream areas along multiple wind corridors. No significant wind blockage can be captured along SE and SSE wind directions.", "hash": "f1135bf9d82ac28d185b0a156d8f5e04", "id": 5}, {"snippet_links": [], "samples": [{"hash": "bR47k9C1z3L", "uri": "/contracts/bR47k9C1z3L#figure-6", "label": "Licence Agreement Concerning Inclusion of Doctoral Thesis in the Institutional Repository of the University of Leiden", "score": 17.0, "published": true}], "size": 1, "snippet": "6: Spectral shift of a molecule line as a function of the log of time. We previously applied different source-drain voltages for 10 minutes and measured the spectral position of the molecule during 10 minutes. We applied a gate voltage Vg=- 50 V during the whole experiment. We fitted the shift with Eq.", "hash": "ba585873967c53f961f89df27f21ba46", "id": 6}, {"snippet_links": [{"key": "figure-7", "type": "clause", "offset": [190, 198]}, {"key": "the-author", "type": "definition", "offset": [325, 335]}, {"key": "figure-8", "type": "definition", "offset": [341, 349]}], "samples": [{"hash": "bkkcrjbw2UG", "uri": "/contracts/bkkcrjbw2UG#figure-6", "label": "Doctoral Thesis", "score": 32.0983107875, "published": true}], "size": 1, "snippet": "A broken cylinder seal shows a row of Egyptian prisoners, and a Persian figure spearing an Egyptian king. (Photo from \u2587\u2587\u2587\u2587\u2587://\u2587\u2587\u2587\u2587\u2587\u2587.\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587.\u2587\u2587\u2587/CRDImages/an/original/DP-16934-022.jpg) 97 Figure 7. \u2587\u2587\u2587\u2587\u2587\u2587 \u2587\u2019\u2587 inscription on the rock face of mount Bisitun (center, above scaffolding) as seen from the ground. (Photograph by the author) 108 Figure 8. A door panel from a wooden naos inscribed with the cartouches of \u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587 \u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587. (Photo from \u2587\u2587\u2587\u2587\u2587://\u2587\u2587\u2587\u2587\u2587\u2587.\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587\u2587.\u2587\u2587\u2587/wikipedia/commons/4/42/Pedubast_II_door.jpg) 134", "hash": "f7648e0bd21373ea9e2cb24cc8e94a5a", "id": 7}, {"snippet_links": [{"key": "major-roadway", "type": "definition", "offset": [204, 217]}, {"key": "a-p", "type": "clause", "offset": [223, 226]}], "samples": [{"hash": "e1YmhIVoDAg", "uri": "/contracts/e1YmhIVoDAg#figure-6", "label": "Distribution Agreement", "score": 20.2692642212, "published": true}], "size": 1, "snippet": "Maximum FEV1/FVC ratio after bronchodilation correlates with roadway proximity in this sample. The maximum FEV1/FVC ratio after bronchodilation was lower in patients who lived less than 300 meters from a major roadway with a p-value of .039 and a r-squared value of 0.023. In patients who lived less than 300 meters from a major roadway, the median value was 93.3 with a 25th percentile value of 83.4 and a 75th percentile value of 98.1. In patients who lived at least 300 meters from a major roadway, the median value was 97.1 with a 25th percentile value of 90.9 and a 75th percentile value of 103.", "hash": "b6cadf6bc342b66de6de5b0069574e5b", "id": 8}, {"snippet_links": [{"key": "group-by", "type": "clause", "offset": [249, 257]}, {"key": "first-generation", "type": "definition", "offset": [1172, 1188]}, {"key": "our-data", "type": "definition", "offset": [1393, 1401]}, {"key": "best-fit", "type": "definition", "offset": [1608, 1616]}, {"key": "standard-deviation", "type": "definition", "offset": [1644, 1662]}, {"key": "distribution-of", "type": "clause", "offset": [1727, 1742]}, {"key": "as-shown", "type": "definition", "offset": [2612, 2620]}, {"key": "figure-9", "type": "definition", "offset": [2624, 2632]}, {"key": "for-example", "type": "clause", "offset": [2678, 2689]}, {"key": "large-enough", "type": "definition", "offset": [3115, 3127]}, {"key": "periods-of", "type": "clause", "offset": [3671, 3681]}, {"key": "other-applications", "type": "definition", "offset": [3749, 3767]}, {"key": "de-novo", "type": "definition", "offset": [3798, 3805]}, {"key": "lack-of-improvement", "type": "clause", "offset": [3917, 3936]}, {"key": "limitations-on", "type": "definition", "offset": [4159, 4173]}, {"key": "own-time", "type": "definition", "offset": [4201, 4209]}, {"key": "number-of", "type": "clause", "offset": [4214, 4223]}, {"key": "the-requirement", "type": "clause", "offset": [5082, 5097]}, {"key": "local-optimization", "type": "clause", "offset": [5461, 5479]}, {"key": "resulting-from", "type": "definition", "offset": [5837, 5851]}], "samples": [{"hash": "8oH8ys9yEpN", "uri": "/contracts/8oH8ys9yEpN#figure-6", "label": "Interactive Evolutionary Algorithms and Data Mining for Drug Design", "score": 19.0342235565, "published": true}], "size": 1, "snippet": "7: The optimization trajectory leading to the best molecule produced by the fragment-based evolution. The numbers under the structures indicate their fitness score. Note that the SO3C2H3-group of the second parent molecule was modified to a SO2C2H3-group by an error in our program. This bug was later fixed. In later fitness comparisons, not much influence on fitness was found by the presence or absence of the extra oxygen atom, so this glitch will probably not have influenced evolution much. To study the differences between the atom-based and fragment-based evolution, we first gathered of each generation the maximum fitness value (the fitness of the \u201cbest\u201d molecule) and the average fitness value. These values are plotted in Figure 6.8. Atom average Atom best Fragment average Fragment best Figure 6.8 shows that both the average fitness and the maximum fitness of the molecules in the population grow as the evolution proceeds. This means that new and better molecules are found, which implies that evolution improves upon pure selection (since pure selection would cause the average and maximum fitnesses of the later generations to approach the maximum of the first generation). However, virtual screening of a large library will also increase maximum fitness, as there is always a probability that a new molecule will improve upon the known compounds. We should therefore analyze our data further to be able to say whether evolution is truly more effective or efficient than random search. Fitting the fitnesses of the randomly generated first generation to a Gaussian (Figure 6.9) results in a best fit with mean value 36.7 and a standard deviation of 4.3. Initial population Final Fragment-based Figure 6.9: The distribution of fitness scores of the molecules of three populations fitted to Gaussians. The populations are the initial population, the atom-based population containing the highest-scoring atom-based molecule (8th atom-based generation), and the fragment-based population containing the highest-scoring fragment-based molecule (the 10th fragment- based generation). The best overall result of evolution (score=76) lies about 9 standard deviations from this average. Therefore the probability that a random search would produce a molecule with this or a higher fitness would be smaller than 10-9%. Scanning 500 molecules and getting the improvement shown in Figure 6.9 suggests that this improvement cannot be the result of a random search only. Therefore evolution seems truly more efficient than random search. We should note hereby that the distribution of the initial population, as shown in Figure 9, is Gaussian-like, but not exactly Gaussian. For example, there are about 5 molecules with a fitness score of about 50, which is more than would be expected from a true Gaussian distribution. Therefore, putting the odds that random search produces a similar improvement as the evolutionary algorithm at 1:1011 may be mathematically questionable. However, the distance between the best fitness found by the EAs and the average and best fitnesses of randomly produced molecules seems large enough to indicate that both of our evolutionary algorithms noticeably improve on random search. While evolution improves molecular structures significantly, the improvement (at least in atom-based evolution) stops quite quickly, around the 8th generation. It is very unlikely that by then the \u2018best possible\u2019 molecule has already been found. What then causes this stagnation? And how could it be prevented in future experiments? The first explanation is that we simply have evaluated too few generations for full optimization. EAs are known to have periods of stasis, and are usually run over many more generations than 10: in other applications of evolutionary algorithms to de novo design molecules are evolved over 50 to 20,000 generations (\u2587\u2587\u2587\u2587\u2587\u2587\u2587, 1998; Vinkers, 2003; \u2587\u2587\u2587\u2587\u2587, 2004). Often, lack of improvement in fitness only means that the population is at some local optimum, and needs to be evolved for a few more generations before better molecules are found. Running the EA for more generations (something we did not do due to limitations on computational time and our own time \u2013 a number of steps such as screening for unusual structures had not yet been automated and required manual intervention) may result in renewed improvement with better molecules found. Our second hypothesis is that the stagnation and deterioration of scores has uncovered an inefficiency in the EA that should be addressed instead of compensated through evaluating more generations. The most likely explanation of the decrease is that our rule that each molecule generated must be \u201cnew\u201d forces the evolution away too quickly from a local optimum; better variants of the best molecule cannot be built anymore since the best molecule has already been \u201cforgotten\u201d. The result is a downhill trend in the fitness scores, the molecular structures going away from the last found optimum, until a novel structure is found which can be optimized again. In our case, it may be that the requirement that each molecule must be new may prevent a true local optimum to be reached since eight generations is too short for full optimization; of any compound, including the best compound, several hundreds of derivatives can be made through atom-based mutation. Therefore, the one or two generations the best molecule had to procreate are not enough for a good \u201cfull\u201d local optimization. Therefore, it would be best to drop the \u201conly new molecules\u201d requirement. Interestingly, the fragment-based evolution does not show a similar deterioration. It is possible that the fragment-evolution has not reached a similar plateau yet, since it reached its peak later than the atom-based evolution. Next to the fitness scores, we compared the molecules resulting from atom-based and fragment-based evolution. The top 5 molecules found by both approaches are shown in Figure 6.10, together with their fitness scores. The most interesting observation to us was that fragment-based evolution seems approximately as good as atom-based evolution. This seems to refute our hypothesis that atom-based evolution should go more smoothly and therefore be better than fragment-based evolution. Why were we wrong? To find an explanation, we studied the different optimization trajectories and the molecules produced in more detail. The evolution trajectories leading to the best molecules found by the atom-based and the fragment-based evolution, respectively, are shown in Figures 6.6 and 6.7.", "hash": "6760d095073f7c59918f5673c28dc5cf", "id": 9}, {"snippet_links": [], "samples": [{"hash": "892pZbhmn2M", "uri": "/contracts/892pZbhmn2M#figure-6", "label": "Grant Agreement", "score": 31.9094257355, "published": true}], "size": 1, "snippet": "Model of the stress-strain-curve in the complete elastic-plastic region", "hash": "533b6a8c463e50588544230f2283f8b4", "id": 10}], "next_curs": "ClESS2oVc35sYXdpbnNpZGVyY29udHJhY3Rzci0LEhZDbGF1c2VTbmlwcGV0R3JvdXBfdjU2IhFmaWd1cmUtNiMwMDAwMDAwYQyiAQJlbhgAIAA=", "clause": {"title": "Figure 6", "size": 9, "parents": [["under-prevailing-e-and-ese-wind", "Under Prevailing E and Ese Wind"]], "children": [], "id": "figure-6", "related": [["technical-evaluation", "TECHNICAL EVALUATION", "TECHNICAL EVALUATION"], ["check-meters", "Check Meters", "Check Meters"], ["job-evaluation", "Job Evaluation", "Job Evaluation"], ["annual-evaluation", "Annual Evaluation", "Annual Evaluation"], ["annual-evaluations", "Annual Evaluations", "Annual Evaluations"]], "related_snippets": [], "updated": "2025-07-24T06:48:57+00:00"}, "json": true, "cursor": ""}}