Typing mineral deposits using their grades and tonnages in an artificial neural network

A test of the ability of a probabilistic neural network to classify deposits into types on the basis of deposit tonnage and average Cu, Mo, Ag, Au, Zn, and Pb grades is conducted. The purpose is to examine whether this type of system might serve as a basis for integrating geoscience information available in large mineral databases to classify sites by deposit type. Benefits of proper classification of many sites in large regions are relatively rapid identification of terranes permissive for deposit types and recognition of specific sites perhaps worthy of exploring further. Total tonnages and average grades of 1,137 well-explored deposits identified in published grade and tonnage models representing 13 deposit types were used to train and test the network. Tonnages were transformed by logarithms and grades by square roots to reduce effects of skewness. All values were scaled by subtracting the variable's mean and dividing by its standard deviation. Half of the deposits were selected randomly to be used in training the probabilistic neural network and the other half were used for independent testing. Tests were performed with a probabilistic neural network employing a Gaussian kernel and separate sigma weights for each class (type) and each variable (grade or tonnage). Deposit types were selected to challenge the neural network. For many types, tonnages or average grades are significantly different from other types, but individual deposits may plot in the grade and tonnage space of more than one type. Porphyry Cu, porphyry Cu-Au, and porphyry Cu-Mo types have similar tonnages and relatively small differences in grades. Redbed Cu deposits typically have tonnages that could be confused with porphyry Cu deposits, also contain Cu and, in some situations, Ag. Cyprus and kuroko massive sulfide types have about the same tonnages. Cu, Zn, Ag, and Au grades. Polymetallic vein, sedimentary exhalative Zn-Pb, and Zn-Pb skarn types contain many of the same metals. Sediment-hosted Au, Comstock Au-Ag, and low-sulfide Au-quartz vein types are principally Au deposits with differing amounts of Ag. Given the intent to test the neural network under the most difficult conditions, an overall 75% agreement between the experts and the neural network is considered excellent. Among the largestclassification errors are skarn Zn-Pb and Cyprus massive sulfide deposits classed by the neuralnetwork as kuroko massive sulfides—24 and 63% error respectively. Other large errors are the classification of 92% of porphyry Cu-Mo as porphyry Cu deposits. Most of the larger classification errors involve 25 or fewer training deposits, suggesting that some errors might be the result of small sample size. About 91% of the gold deposit types were classed properly and 98% of porphyry Cu deposits were classes as some type of porphyry Cu deposit. An experienced economic geologist would not make many of the classification errors that were made by the neural network because the geologic settings of deposits would be used to reduce errors. In a separate test, the probabilistic neural network correctly classed 93% of 336 deposits in eight deposit types when trained with presence or absence of 58 minerals and six generalized rock types. The overall success rate of the probabilistic neural network when trained on tonnage and average grades would probably be more than 90% with additional information on the presence of a few rock types.

Potent T cell response to a class I-binding 13-mer viral epitope and the influence of HLA micropolymorphism in controlling epitope length

The BZLF1 antigen of Epstein-Barr virus includes three overlapping sequences of different lengths that conform to the binding motif of human leukocyte antigen (HLA) B*3501. These 9-mer ((56)LPOGQLTAy(64)), 11-mer ((54)EPLPQGQLTAy(64)), and 13-mer ((52)LPEPLPQGQLTAY(64)) peptides all bound well to B*3501; however, the CTL response in individuals expressing this HILA allele was directed strongly and exclusively towards the 11-mer peptide. In contrast, EBV-exposed donors expressing HLA B*3503 showed no significant CTL response to these peptides because the single amino acid difference between B*3501 and B*3503 within the F pocket inhibited HLA binding by these peptides. The extraordinarily long 13-mer peptide was the target for the CTL response in individuals expressing B*3508, which differs from B*3501 at a single position within the D pocket (B*3501, 156 Leucine; B*3508, 156 Arginine). This minor difference was shown to enhance binding of the 13-mer peptide, presumably through a stabilizing interaction between the negatively charged glutamate at position 3 of the peptide and the positively charged arginine at HLA position 156. The 13-mer epitope defined in this study represents the longest class I-binding viral epitope identified to date as a minimal determinant. Furthermore, the potency of the response indicates that peptides of this length do not present a major structural barrier to CTL recognition.

Development of lipid-core-peptide (LCP) based vaccines for the prevention of the group A streptococcal (GAS) infection

Traditional vaccines consisting of whole attenuated micro-organisms. or microbial components administered with adjuvant, have been demonstrated as one of the most cost-effective and successful public health interventions. Their use in large scale immunisation programs has lead to the eradication of smallpox, reduced morbidity and mortality from many once common diseases, and reduced strain on health services. However, problems associated with these vaccines including risk of infection. adverse effects, and the requirement for refrigerated transport and storage have led to the investigation of alternative vaccine technologies. Peptide vaccines, consisting of either whole proteins or individual peptide epitopes, have attracted much interest, as they may be synthesised to high purity and induce highly specific immune responses. However, problems including difficulties stimulating long lasting immunity. and population MHC diversity necessitating multiepitopic vaccines and/or HLA tissue typing of patients complicate their development. Furthermore, toxic adjuvants are necessary to render them immunogenic. and as such non-toxic human-compatible adjuvants need to be developed. Lipidation has been demonstrated as a human compatible adjuvant for peptide vaccines. The lipid-core-peptide (LCP) system. incorporating lipid adjuvant, carrier, and peptide epitopes, exhibits promise as a lipid-based peptide vaccine adjuvant. The studies reviewed herein investigate the use of the LCP system for developing vaccines to protect against group A streptococcal (GAS) infection. The studies demonstrate that LCP-based GAS vaccines are capable of inducing high-titres of antigen specific IgG antibodies. Furthermore. mice immunised with an LCP-based GAS vaccine were protected against challenge with 8830 strain GAS.

Antigenic and genetic typing of Whataroa viruses in Australia

We recently characterized three novel alphaviruses isolated from mosquitoes captured in New South Wales, Australia. Initial cross-neutralization studies revealed antigenic similarity to the Sindbis virus (SINV)-like Whataroa virus (WHAV), heretofore found only in New Zealand. Nucleotide sequence analysis showed that the WHAV-Iike viruses shared >99% nucleotide sequence similarity with each other, and 96-97% similarity with prototype WHAV. Enzyme-linked immunosorbent assay reactions of a panel of monoclonal antibodies to SINV showed that the novel WHAV-Iike viruses displayed identical binding patterns and were antigenically distinct from all SINV isolates examined. Although these viruses displayed a similar binding pattern to prototype WHAV, three monoclonal antibodies discriminated them from the New Zealand virus. Our results suggest that these novel alphaviruses are antigenic variants of WHAV and represent the first reported isolations of this virus from outside New Zealand. The monoclonal antibodies used in this study will be useful for typing new SINV and SINV-like isolates.

MULTIPRED: A computational system for prediction of promiscuous HLA binding peptides

MULTIPRED is a web-based computational system for the prediction of peptide binding to multiple molecules ( proteins) belonging to human leukocyte antigens (HLA) class I A2, A3 and class II DR supertypes. It uses hidden Markov models and artificial neural network methods as predictive engines. A novel data representation method enables MULTIPRED to predict peptides that promiscuously bind multiple HLA alleles within one HLA supertype. Extensive testing was performed for validation of the prediction models. Testing results show that MULTIPRED is both sensitive and specific and it has good predictive ability ( area under the receiver operating characteristic curve A(ROC) > 0.80). MULTIPRED can be used for the mapping of promiscuous T-cell epitopes as well as the regions of high concentration of these targets termed T-cell epitope hotspots. MULTIPRED is available at http:// antigen.i2r.a-star.edu.sg/ multipred/.

A population of HLA-DR+ immature cells accumulates in the blood dendritic cell compartment of patients with different types of cancer

Dendritic cell (DC) defects are an important component of immunosuppression in cancer. Here, we assessed whether cancer could affect circulating DC populations and its correlation with tumor progression. The blood DC compartment was evaluated in 136 patients with breast cancer, prostate cancer, and malignant glioma. Phenotypic, quantitative, and functional analyses were performed at various stages of disease. Patients had significantly fewer circulating myeloid (CD11c(+)) and plasmacytoid (CD123(+)) DC, and a concurrent accumulation of CD11c(-)CD123(-) immature cells that expressed high levels of HLA-DR+ immature cells (DR+IC). Although DR+IC exhibited a limited expression of markers ascribed to mature hematopoietic lineages, expression of HLA-DR, CD40, and CD86 suggested a role as antigen-presenting cells. Nevertheless, DR+IC had reduced capacity to capture antigens and elicited poor proliferation and interferon-gamma secretion by T-lymphocytes. Importantly, increased numbers of DR+IC correlated with disease status. Patients with metastatic breast cancer showed a larger number of DR+IC in the circulation than patients with local/nodal disease. Similarly, in patients with fully resected glioma, the proportion of DR+IC in the blood increased when evaluation indicated tumor recurrence. Reduction of blood DC correlating with accumulation of a population of immature cells with poor immunologic function may be associated with increased immunodeficiency observed in cancer.

Extreme learning machine for predicting HLA-peptide binding

Machine learning techniques have been recognized as powerful tools for learning from data. One of the most popular learning techniques, the Back-Propagation (BP) Artificial Neural Networks, can be used as a computer model to predict peptides binding to the Human Leukocyte Antigens (HLA). The major advantage of computational screening is that it reduces the number of wet-lab experiments that need to be performed, significantly reducing the cost and time. A recently developed method, Extreme Learning Machine (ELM), which has superior properties over BP has been investigated to accomplish such tasks. In our work, we found that the ELM is as good as, if not better than, the BP in term of time complexity, accuracy deviations across experiments, and most importantly - prevention from over-fitting for prediction of peptide binding to HLA.

Prediction of HLA-DQ3.2 beta ligands: evidence of multiple registers in class II binding peptides

Motivation: While processing of MHC class II antigens for presentation to helper T-cells is essential for normal immune response, it is also implicated in the pathogenesis of autoimmune disorders and hypersensitivity reactions. Sequence-based computational techniques for predicting HLA-DQ binding peptides have encountered limited success, with few prediction techniques developed using three-dimensional models. Methods: We describe a structure-based prediction model for modeling peptide-DQ3.2 beta complexes. We have developed a rapid and accurate protocol for docking candidate peptides into the DQ3.2 beta receptor and a scoring function to discriminate binders from the background. The scoring function was rigorously trained, tested and validated using experimentally verified DQ3.2 beta binding and non-binding peptides obtained from biochemical and functional studies. Results: Our model predicts DQ3.2 beta binding peptides with high accuracy [area under the receiver operating characteristic (ROC) curve A(ROC) > 0.90], compared with experimental data. We investigated the binding patterns of DQ3.2 beta peptides and illustrate that several registers exist within a candidate binding peptide. Further analysis reveals that peptides with multiple registers occur predominantly for high-affinity binders.

HLA and Genomewide Allele Sharing in Dizygotic Twins

Gametic selection during fertilization or the effects of specific genotypes on the viability of embryos may cause a skewed transmission of chromosomes to surviving offspring. A recent analysis of transmission distortion in humans reported significant excess sharing among full siblings. Dizygotic (DZ) twin pairs are a special case of the simultaneous survival of two genotypes, and there have been reports of DZ pairs with excess allele sharing around the HLA locus, a candidate locus for embryo survival. We performed an allele-sharing study of 1,592 DZ twin pairs from two independent Australian cohorts, of which 1,561 pairs were informative for linkage on chromosome 6. We also analyzed allele sharing in 336 DZ twin pairs from The Netherlands. We found no evidence of excess allele sharing, either at the HLA locus or in the rest of the genome. In contrast, we found evidence of a small but significant (P = .003 for the Australian sample) genomewide deficit in the proportion of two alleles shared identical by descent among DZ twin pairs. We reconciled conflicting evidence in the literature for excess genomewide allele sharing by performing a simulation study that shows how undetected genotyping errors can lead to an apparent deficit or excess of allele sharing among sibling pairs, dependent on whether parental genotypes are known. Our results imply that gene-mapping studies based on affected sibling pairs that include DZ pairs will not suffer from false-positive results due to loci involved in embryo survival.

Studies of HLA associations in male and female patients with Guillain-Barre syndrome (GBS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP)

HLA associations are found to differ with the gender of the patient in some autoimmune diseases. Here we have investigated whether there are gender-related HLA associations in Guillain-Barre syndrome (GBS) and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), both of which occur more frequently in male patients than in females. In GBS, no particular HLA associations were noted, except for a slight negative association in both males and females for carriage of HLA-DR5. In CIDP, the gene frequency and the frequency of individuals positive for HLA-DR2 were greater in female patients than female controls, although this was statistically significant only for the gene frequency. Furthermore more female CIDP patients were homozygous for DR2, than male CIDP patients, or male or female controls and patients with GBS. This suggests that sex-related factors may interact with the risk associated with carriage of HLA-DR2 for development of CIDP. (c) 2006 Published by Elsevier B.V.