Feature engineered embeddings for classification of molecular data.

The classification of molecules is of particular importance to the drug discovery process and several other use cases. Data in this domain can be partitioned into structural and sequence/text data. Several techniques such as deep learning are able to classify molecules and predict their functions using both types of data. Molecular structure and encoded chemical information are sufficient to classify a characteristic of a molecule. However, the use of a molecule's structural information typically requires large amounts of computational power with deep learning models that take a long time to train. In this study, we present an alternative approach to molecule classification that addresses the limitations of other techniques. This approach uses natural language processing techniques in the form of count vectorisation, term frequency-inverse document frequency, word2vec and Latent Dirichlet Allocation to feature engineer molecular text data. Through this approach, we aim to make a robust and easily reproducible embedding that is fast to implement and solely dependent on chemical (text) data such as the sequence of a protein. Further, we investigate the usefulness of these embeddings for machine learning models. We apply the techniques to two different types of molecular text data: FASTA sequence data and Simplified Molecular Input Line Entry Specification data. We show that these embeddings provide excellent performance for classification.

Metric evaluation of the anterior nasal spine to estimate sex and population group in South African individuals.

INTRODUCTION: The anterior nasal spine is a pointed, midline projection of the maxilla. This bony structure dictates the overlying soft tissues providing the phenotypic features of the nose and upper lip and determines the differences in the mid-face morphology. Little data is available on the metric features of the Anterior nasal spine (ANS). This study aimed to perform metric evaluations of the ANS of white and black South African males and females to ascertain if morphological variations exist and if the differences are viable for the use in sex and population identification. MATERIALS AND METHODS: The sample included 100 CBCT images for each population and sex group. Linear and angular measurements of the ANS were recorded in both the sagittal and axial planes. RESULTS: Classification decision trees (pruned) were fitted to ascertain the relationship between population group, sex and the ANS measurements including and excluding age. For population group, all the ANS measurements were statistically significant for females but in males, all the ANS measurements were significant when performed individually. However, when fitted to the classification tree, Sagittal 2 did not show any statistical significance. When considering sex, only 2 of the ANS measurements (Sagittal 2 and Axial 1) were found to be significant. The results did not differ significantly when comparing the decision trees including and excluding age. CONCLUSIONS: White South African individuals presented with a longer ANS that produced a more acute angle whereas black South African individuals presented with a shorter ANS and a more obtuse angle. Additionally, males presented with a longer ANS compared to females. ANS measurements were found to be more relevant for population discernment than for sex.

Investigation of mandibular second molar root and canal morphology in a Black South African population using cone-beam computed tomography and two classification systems.

PURPOSE: To investigate the root and canal morphology of mandibular second molars, including anatomical variations, in a Black South African population using two classification systems. METHODS: Cone-beam computed tomography images of 386 teeth were evaluated. The number of roots and canal configurations were categorized using the classifications of Vertucci (with modifications by Sert and Bayirli) and Ahmed et al. (including the modified Melton classification). Relationships between variables were assessed using Fisher's exact test (P < 0.05). RESULTS: Mandibular second molars were predominantly two-rooted (91.7%). The majority of teeth had three canals including Type IV mesial (41.2%) and Type I (75.1%) distal (Ahmed et al. configuration: 2MDM M2 D1). Three or more canals were present in almost one-fifth of mesial (n = 72/386, 18.6%) and a small number of distal (n = 21/386, 5.4%) roots. Males displayed additional canals more frequently in the distal root (P = 0.02). C-shaped anatomy was found in 5.7%. Among the C-shapes, the most common was Melton's Type III in the coronal (50%), middle (81%) and apical (72%) thirds. CONCLUSION: In this population, mandibular second molar teeth exhibited diverse morphology, which would have clinical significance for endodontic practitioners. The Ahmed et al. classification provided a better description than the Vertucci classification.

Root and canal morphology of maxillary second molars in a Black South African subpopulation using cone-beam computed tomography and two classifications.

This study investigated the root and canal morphology of maxillary second molars in a Black South African population using high-resolution cone-beam computed tomography. In total, 386 maxillary second molar teeth were evaluated and described according to the classifications of Vertucci (with the additions of Sert and Bayirli), as well as Ahmed et al. Root number, fusions and canal morphology were recorded. Relationships between morphology, sex and age were assessed using Fisher's exact test (p < 0.05). Maxillary second molars were predominantly three-rooted (96.1%), with root fusion observed in 14%. No relationships between sex, age and canal morphology were found. Additional canals were observed in 67.4% of mesiobuccal roots. The Ahmed et al. classification provided a better overall description of the morphology. In this population, maxillary second molar teeth exhibited a diversity of root and canal morphology. Complex variations may render endodontic management more difficult and increase the possibility of missed anatomy.

A study of mandibular premolar root and canal morphology in a Black South African population using cone-beam computed tomography and two classification systems.

PURPOSE: An investigation of the configurations of mandibular premolar roots and canals in a population of Black South Africans. METHODS: Cone-beam computed tomography analysis of 772 mandibular premolars was performed, and the premolars were classified according to the systems proposed by Vertucci and Ahmed et al. Root number, canal morphology, age, and sex were recorded. Fisher's exact test was used to determine relationships based on age and sex (P < 0.05). RESULTS: Single roots were seen in the majority of mandibular premolars (97.1%). Single canal configurations (i.e., Vertucci Type 1/Ahmed et al. 1MP1) were observed in 48.5% of first and 81.3% of second mandibular premolars. Mandibular first premolars demonstrated multiple canals in more than half of the sample (51.5%), and C-shaped morphology in more than one-tenth (11.1%). A relationship between sex and the presence of radicular grooves was demonstrated (P = 0.049), males being more likely to demonstrate this feature (P = 0.051). Multiple canals in mandibular first premolars also showed a relationship with sex (P = 0.005), a male predilection being evident (P = 0.007). The Ahmed et al. system proved superior to the Vertucci classification for reporting complex configurations and anatomical variations, although a greater number of unique categories were created. CONCLUSION: Diverse mandibular premolar root and canal morphology was observed in the studied population. Clinicians must be aware of common morphological features as well as possible anatomical variations in mandibular premolars, as failure to treat complete root canal systems may negatively impact endodontic treatment outcomes.

Home Range and Movement Patterns of Reintroduced White Lions (Panthera leo melanochaita) in the Kruger to Canyons Biosphere Reserve, South Africa.

White lions are a colour variant of the African lion Panthera leo melanochaita and disappeared from the wild due to anthropogenic factors until their reintroduction to the Greater Kruger Park Region of South Africa in 2006. Natural home range behaviour is an index of reintroduction success. Therefore, the home range and movement of a pride of reintroduced white lions and a constructed pride consisting of reintroduced white lions and translocated wild tawny lionesses in small, fenced reserves was assessed. GPS data from collared adults were collected for the white lion pride between 2010-2011 and 2018-2020 for the constructed pride. Home ranges were estimated using kernel density estimation and minimum convex polygon, with minimum daily distance tested for differences between sex, season, and pride. Home ranges were small and average daily movements restricted for both prides (white lion pride: 5.41 km2 and 10.44 ± 4.82 km; constructed pride: 5.50 km2, 11.37 ± 4.72 km) due to the small reserve size of 7 km2. There was no difference between prides for annual and seasonal home range size, male and female home ranges, minimum daily distance travelled, or habitat selection. White lions from both prides established territories and displayed natural home ranging behaviour, suggesting that their reintroduction was successful, in the absence of anthropogenic threats.

Root and canal morphology of the permanent anterior dentition in a Black South African population using cone-beam computed tomography and two classification systems.

PURPOSE: This study investigated the root and canal morphology of the permanent anterior dentition in a Black South African population. METHODS: In total 2,343 permanent anterior teeth were evaluated using cone-beam computed tomography images and described according to the Vertucci and Ahmed et al. classifications. Demographic information, root number and internal canal morphology were recorded. Age and sex associations were made using Fisher's exact test (P < 0.05). RESULTS: All anterior teeth were single-rooted (100%). The most frequent configuration in maxillary anteriors was a single canal (Vertucci Type I/1MXA1). In mandibular anteriors, single canal configurations (Type I/1MDA1) predominated, however 36.6% of mandibular central and 33.9% of lateral incisors displayed two canals. Older subjects (>40 years) demonstrated single canals more than younger subjects (P = 0.0004). Females displayed variations more compared to males (P = 0.002). The two classifications were found to be comparable for evaluation of permanent anterior teeth. Complex configurations were better described using the newer system. CONCLUSION: The permanent anterior teeth of this population exhibited exclusively single roots and diverse internal characteristics. Clinicians should be aware of anatomical variations, particularly in mandibular incisors, as these may result in adverse endodontic outcomes if not taken into consideration by the treating practitioner.

The association between weather and crime in a township setting in South Africa.

The association between various meteorological parameters and crime is well-established in developed contexts. In contrast in this study, we investigated the association between three weather parameters (temperature, relative humidity and rainfall) and three categories of crime in the developing township of Khayelitsha, in the Western Cape Province of South Africa. Distributed lag non-linear modelling was used to identify temporal relationships between temperature, relative humidity and rainfall, and violent, property and sexual crime over a 10-year period (2006-2016). We found hot days (defined as [Formula: see text] 25 °C) increased the cumulative relative risk of violent crime by up to 32% but were also found to be associated with a lagged increase in violent crime for at least a week thereafter. On very cold days (defined as [Formula: see text]), the cumulative relative risk of property crime increased by up to 50% whereas on very rainy days (defined as [Formula: see text]) the risk of property crime surprisingly increased by 40%. These findings provide some additional evidence for the relationship between the atmospheric environment and human behaviour in a developing context.

LULU operators and discrete pulse transform for multi-dimensional arrays.

The LULU operators for sequences, that is L(n), U(n) and their compositions, are extended to multidimensional arrays in a way which preserves their essential properties, e.g., consistent separation, total variation and shape preservation. The power of the operators is demonstrated by deriving the discrete pulse transform (DPT), which is a hierarchical decomposition of the input array into pulses. Similar to its 1-D counterpart this transform satisfies a basic consistency property.