A deep learning approach for the forensic evaluation of sexual assault

Despite the existence of patterns able to discriminate between consensual and non-consensual intercourse, the relevance of genital lesions in the corroboration of a legal rape complaint is currently under debate in many countries. The testimony of the physicians when assessing these lesions has been questioned in court due to several factors (e.g., a lack of comprehensive knowledge of lesions, wide spectrum of background area, among others). Therefore, it is relevant to provide automated tools to support the decision process in an objective manner. In this work, we evaluate the performance of state-of-the-art deep learning architectures for the forensic assessment of sexual assault. We propose a deep architecture and learning strategy to tackle the class imbalance on deep learning using ranking. The proposed methodologies achieved the best results when compared with handcrafted feature engineering and with other deep architectures.     

Analysis on splice, microbending, macrobending, and Rayleigh losses in GeO2-doped dispersion-shifted single-mode fibers

Independent of the actual index profile of a single-cladded dispersion-shifted single-mode fiber, it is shown that the loss components given in the headline at the zero dispersion wavelength λ0accurately can be obtained by simple expressions that only depend on the Laplace spot size value at λ0. Withlambda_{0} = 1.55 mum, these relations are used to estimate and minimize the total loss in single-cladded dispersion-shifted fibers with arbitrary core-index profiles.     

Type-1 plasminogen-activator inhibitor -- conformational differences between latent, active, reactive-centre-cleaved and plasminogen-activator-complexed forms, as probed by proteolytic susceptibility

We have analysed the susceptibility of latent, active, reactive-centre-cleaved and plasminogen-activator-complexed type-1 plasminogen-activator inhibitor (PAI-1) to the non-target proteinases trypsin, endoproteinase Asp-N, proteinase K and subtilisin. This analysis has allowed us to detect conformational differences between the different forms of PAI-1 outside the reactive-centre loop and beta-sheet A. Proteinase-hypersensitive sites were clustered in three regions. Firstly, susceptibility was observed in the region around alpha-helix E, beta-strand 1A, and the flanking loops, which are believed to form flexible joints during movements of beta-sheet A. Secondly, hypersensitive sites were observed in the loop between alpha-helix I and beta-strand 5A. Thirdly, the gate region, encompassing beta-strands 3C and 4C, was highly susceptible to trypsin in latent PAI-1, but not in the other conformations. The digestion patterns differed among all four forms of PAI-1, indicating that each represents a unique conformation. The differential proteolytic susceptibility of the flexible-joint region may be coupled to the differential affinity to vitronectin, binding in the same region. The analysis also allowed detection of conformational differences between reactive-centre-cleaved forms produced under different solvent conditions. The digestion pattern of plasminogen-activator-complexed PAI-1 was different from that of active PAI-1, but indistinguishable from that of one of the reactive-centre-cleaved forms, as the complexed and this particular cleaved PAI-1 were completely resistant to all the non-target proteinases tested. This observation is in agreement with the notion that complex formation involves reactive-centre cleavage and a large degree of insertion of the reactive-centre loop into beta-sheet A. Our analysis has allowed the identification of some flexible regions that appear to be implicated in the conformational changes during the movements of beta-sheet A and during the inhibitory reaction of serpins with their target proteinases.     

Rapid disappearance of zinc positive terminals in focal brain ischemia

The study was undertaken to determine if the levels of vesicular zinc in neuronal terminals would decrease in response to focal brain ischemia. The middle cerebral artery was occluded distal to the striatal branches in male spontaneously hypertensive rats. At 7, 15, 30, 45, 60, 90, 120 min; 3, 6, 12, 24, 48 h and 7 days later the animals were sacrificed and the brains were stained for zinc-sulfides, cell bodies and AChE-positive cholinergic fibers. The density of zinc positive terminals significantly decreased in the neocortical ischemic zone 7 min after middle cerebral artery occlusion (MCAO). In the neocortical layers II and III most zinc positive neuronal terminals disappeared at 7 min after MCAO whereas the zinc positive terminals in layers V and VI remained positive at least 2 h. Beginning at 1 h after MCAO and progressing to 24 h a significant decrease in the density of zinc positive terminals was observed in the dorsolateral striatum, and ventrobasal thalamic nucleus, both major projection areas of the sensorimotor cortex. The disappearance of zinc positive neuronal terminals in the ischemic neocortex and related areas, is most likely due to a neuronal release of vesicular zinc in response to hypoxia. The high extracellular concentration of zinc is thought to be both neuroprotective by blocking the NMDA receptor and neurotoxic by activating neuronal influx of Ca2+ through voltage gated calcium channels. It seems evident that the latter effect of zinc is contributing to the neuronal death in focal brain ischemia.     

The effect of age-dependent host mortality on the dynamics of an endemic disease

Using asymptotic expansions in the ratio between the duration of infection and host lifetime, equilibrium conditions are analyzed for an SIR-type epidemic model with age-dependent mortality and age-independent disease transmission. Disease incidence at equilibrium depends on the distribution of lifetimes. Incidence is maximal if host life span is fixed and, for vanishing higher moments, it decreases with increasing variance of the distribution. The spectrum of the linearization about the endemic equilibrium has two dominant components, one near 0 and one with a large imaginary part. All roots of the characteristic equation have a negative real part so the model is always stable. The roots with a large imaginary part dominate in most cases, indicating that the approach to equilibrium will be through slowly damped oscillations.     

Improved Endurance Limit of Zirconia Ceramics by Overloading

The influence of initial overloads upon the subsequent fatigue crack growth in transformation-toughened ceramics (TTCs) is investigated. It is revealed that initial overloading can be beneficial to these ceramics. In particular, it is shown that an overload will facilitate crack arrest in a low transformation strength material which would otherwise be prone to cyclic fatigue degradation and eventual failure. Based on the results of this investigation, a proof test may be designed to screen out TTC components which do not attain a prescribed endurance limit.     

Sputtering of Surfaces of the Solid Hydrogens

Sputtering of the solid hydrogens by electrons and ions exhibits features that may be related to quantum properties of these solids, i.e. a drastic enhancement of the yield for electron–bombarded thick deuterium films and a thermal peak at low ejection energies in the energy distribution of the sputtered particles.     

Loop Protein Engineering for Improved Transglycosylation Activity of a β‐N‐Acetylhexosaminidase

Certain enzymes of the glycoside hydrolase family 20 (GH20) exert transglycosylation activity and catalyze the transfer of β‐N‐acetylglucosamine (GlcNAc) from a chitobiose donor to lactose to produce lacto‐N‐triose II (LNT2), a key human milk oligosaccharide backbone moiety. The present work is aimed at increasing the transglycosylation activity of two selected hexosaminidases, HEX1 and HEX2, to synthesize LNT2 from lactose and chitobiose. Peptide pattern recognition analysis was used to categorize all GH20 proteins in subgroups. On this basis, we identified a series of proteins related to HEX1 and HEX2. By sequence alignment, four additional loop sequences were identified that were not present in HEX1 and HEX2. Insertion of these loop sequences into the wild‐type sequences induced increased transglycosylation activity for three out of eight mutants. The best mutant, HEX1_GTEPG, had a transglycosylation yield of LNT2 on the donor that was nine times higher than that of the wild‐type enzyme. Homology modeling of the enzymes revealed that the loop insertion produced a more shielded substrate‐binding pocket. This shielding is suggested to explain the reduced hydrolytic activity, which in turn resulted in the increased transglycosylation activity of HEX1_GTEPG.     

Lack of an association between delayed memory and hippocampal and temporal lobe size in patients with schizophrenia and healthy controls

The purpose of the present study was to investigate putative neural substrates of long-term (delayed) memory in schizophrenia and young healthy controls. Ten "low" and 10 "high" memory patients were selected from a large sample of DSM-III-R diagnosed schizophrenia spectrum patients, based on composite verbal and nonverbal delayed recall memory scores. Ten "low" and 9 "high" memory individuals were also selected from a larger sample of young healthy controls. Magnetic resonance imaging scans were acquired on a 1.5-T GE Signa scanner using a SPGR sequence (repetition time = 24 msec, echo time = 5 msec). Hippocampal volumes were computed from manual tracings (intraclass correlation = .96), and temporal lobe and whole brain tissue volumes were obtained using a semiautomated technique. In both the patient sample and controls, there was no significant relationship between delayed memory ability and hippocampal, temporal lobe, or whole brain volume. The integration of results from this study, and from studies on normal aging and Alzheimer's disease, supports a model suggesting that hippocampal size may be an indicator of long-term memory ability, but only when hippocampal measures reflect aging and degenerative hippocampal atrophy. If the hippocampal measures reflect individual differences in hippocampal size prior to the onset of hippocampal atrophy, then hippocampal size does not appear to predict long-term memory ability.