Using Manifold Learning for Atlas Selection in Multi-Atlas Segmentation

Multi-atlas segmentation has been widely used to segment various anatomical structures. The success of this technique partly relies on the selection of atlases that are best mapped to a new target image after registration. Recently, manifold learning has been proposed as a method for atlas selection. Each manifold learning technique seeks to optimize a unique objective function. Therefore, different techniques produce different embeddings even when applied to the same data set. Previous studies used a single technique in their method and gave no reason for the choice of the manifold learning technique employed nor the theoretical grounds for the choice of the manifold parameters. In this study, we compare side-by-side the results given by 3 manifold learning techniques (Isomap, Laplacian Eigenmaps and Locally Linear Embedding) on the same data set. We assess the ability of those 3 different techniques to select the best atlases to combine in the framework of multi-atlas segmentation. First, a leave-one-out experiment is used to optimize our method on a set of 110 manually segmented atlases of hippocampi and find the manifold learning technique and associated manifold parameters that give the best segmentation accuracy. Then, the optimal parameters are used to automatically segment 30 subjects from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). For our dataset, the selection of atlases with Locally Linear Embedding gives the best results. Our findings show that selection of atlases with manifold learning leads to segmentation accuracy close to or significantly higher than the state-of-the-art method and that accuracy can be increased by fine tuning the manifold learning process.     

Cost-Effectiveness Analysis of the Bivalent and Quadrivalent Human Papillomavirus Vaccines from a Societal Perspective in Colombia

Objective

To compare costs and effectiveness of three strategies used against cervical cancer (CC) and genital warts: (i) Screening for CC; (ii) Bivalent Human Papillomavirus (HPV) 16/18 vaccine added to screening; (iii) Quadrivalent HPV 6/11/16/18 vaccine added to screening.

Methods

A Markov model was designed in order to simulate the natural history of the disease from 12 years of age (vaccination) until death. Transition probabilities were selected or adjusted to match the HPV infection profile in Colombia. A systematic review was undertaken in order to derive efficacy values for the two vaccines as well as for the operational characteristics of the cytology test. The societal perspective was used. Effectiveness was measured in number of averted Disability Adjusted Life Years (DALYS).

Results

At commercial prices reported for 2010 the two vaccines were shown to be non-cost-effective alternatives when compared with the existing screening strategy. Sensitivity analyses showed that results are affected by the cost of vaccines and their efficacy values, making it difficult to determine with certainty which of the two vaccines has the best cost-effectiveness profile. To be ‘cost-effective’ vaccines should cost between 141 and 147 USD (Unite States Dollars) per vaccinated girl at the most. But at lower prices such as those recommended by WHO or the price of other vaccines in Colombia, HPV vaccination could be considered very cost-effective.

Conclusions

HPV vaccination could be a convenient alternative for the prevention of CC in Colombia. However, the price of the vaccine should be lower for this vaccination strategy to be cost-effective. It is also important to take into consideration the willingness to pay, budgetary impact, and program implications, in order to determine the relevance of a vaccination program in this country, as well as which vaccine should be selected for use in the program.     

High-Throughput Sequencing of a South American Amerindian

The emergence of next-generation sequencing technologies allowed access to the vast amounts of information that are contained in the human genome. This information has contributed to the understanding of individual and population-based variability and improved the understanding of the evolutionary history of different human groups. However, the genome of a representative of the Amerindian populations had not been previously sequenced. Thus, the genome of an individual from a South American tribe was completely sequenced to further the understanding of the genetic variability of Amerindians. A total of 36.8 giga base pairs (Gbp) were sequenced and aligned with the human genome. These Gbp corresponded to 95.92% of the human genome with an estimated miscall rate of 0.0035 per sequenced bp. The data obtained from the alignment were used for SNP (single-nucleotide) and INDEL (insertion-deletion) calling, which resulted in the identification of 502,017 polymorphisms, of which 32,275 were potentially new high-confidence SNPs and 33,795 new INDELs, specific of South Native American populations. The authenticity of the sample as a member of the South Native American populations was confirmed through the analysis of the uniparental (maternal and paternal) lineages. The autosomal comparison distinguished the investigated sample from others continental populations and revealed a close relation to the Eastern Asian populations and Aboriginal Australian. Although, the findings did not discard the classical model of America settlement; it brought new insides to the understanding of the human population history. The present study indicates a remarkable genetic variability in human populations that must still be identified and contributes to the understanding of the genetic variability of South Native American populations and of the human populations history.     

Effects of Highly Conserved Major Histocompatibility Complex (MHC) Extended Haplotypes on Iron and Low CD8+ T Lymphocyte Phenotypes in HFE C282Y Homozygous Hemochromatosis Patients from Three Geographically Distant Areas

Hereditary Hemochromatosis (HH) is a recessively inherited disorder of iron overload occurring commonly in subjects homozygous for the C282Y mutation in HFE gene localized on chromosome 6p21.3 in linkage disequilibrium with the human leukocyte antigen (HLA)-A locus. Although its genetic homogeneity, the phenotypic expression is variable suggesting the presence of modifying factors. One such genetic factor, a SNP microhaplotype named A-A-T, was recently found to be associated with a more severe phenotype and also with low CD8⁺T-lymphocyte numbers. The present study aimed to test whether the predictive value of the A-A-T microhaplotype remained in other population settings. In this study of 304 HH patients from 3 geographically distant populations (Porto, Portugal 65; Alabama, USA 57; Nord-Trøndelag, Norway 182), the extended haplotypes involving A-A-T were studied in 608 chromosomes and the CD8⁺ T-lymphocyte numbers were determined in all subjects. Patients from Porto had a more severe phenotype than those from other settings. Patients with A-A-T seemed on average to have greater iron stores (p = 0.021), but significant differences were not confirmed in the 3 separate populations. Low CD8⁺ T-lymphocytes were associated with HLA-A*03-A-A-T in Porto and Alabama patients but not in the greater series from Nord-Trøndelag. Although A-A-T may signal a more severe iron phenotype, this study was unable to prove such an association in all population settings, precluding its use as a universal predictive marker of iron overload in HH. Interestingly, the association between A-A-T and CD8⁺ T-lymphocytes, which was confirmed in Porto and Alabama patients, was not observed in Nord-Trøndelag patients, showing that common HLA haplotypes like A*01–B*08 or A*03–B*07 segregating with HFE/C282Y in the three populations may carry different messages. These findings further strengthen the relevance of HH as a good disease model to search for novel candidate loci associated with the genetic transmission of CD8⁺ T-lymphocyte numbers.     

Blood sampling by puncture in the cavernous sinus from hunted wild boar

A new method of sampling based on the extraction of blood from the cavernous sinus of the dura mater has been assessed in hunted wild boar. Blood from 139 animals was obtained by two different extraction methods: the harvesting from thoracic cavity (TC) and intracavernous venipuncture (IV). Sera obtained by the IV method had higher volume (mean 2.85 vs 1.85 ml), were less hemolytic (mean absorbance at 450 nm: 1.01 vs 2.41 nm). A higher number of samples and a higher proportion of sera collected by IV (90.6 %) compared to those obtained using the TC method (78.4 %), could be analyzed against Aujeszky’s disease using blocking ELISA. No statistically significant differences in seroprevalences between samples obtained using both extraction methods were observed. The results obtained indicate that the IV is an easy, fast, reliable, clean, and safe method to collect blood samples from hunted wild boar, proving a real alternative to the traditional collection method.     

The Genetic Basis of Escherichia coli Pathoadaptation to Macrophages

Antagonistic interactions are likely important driving forces of the evolutionary process underlying bacterial genome complexity and diversity. We hypothesized that the ability of evolved bacteria to escape specific components of host innate immunity, such as phagocytosis and killing by macrophages (MΦ), is a critical trait relevant in the acquisition of bacterial virulence. Here, we used a combination of experimental evolution, phenotypic characterization, genome sequencing and mathematical modeling to address how fast, and through how many adaptive steps, a commensal Escherichia coli (E. coli) acquire this virulence trait. We show that when maintained in vitro under the selective pressure of host MΦ commensal E. coli can evolve, in less than 500 generations, virulent clones that escape phagocytosis and MΦ killing in vitro, while increasing their pathogenicity in vivo, as assessed in mice. This pathoadaptive process is driven by a mechanism involving the insertion of a single transposable element into the promoter region of the E. coli yrfF gene. Moreover, transposition of the IS186 element into the promoter of Lon gene, encoding an ATP-dependent serine protease, is likely to accelerate this pathoadaptive process. Competition between clones carrying distinct beneficial mutations dominates the dynamics of the pathoadaptive process, as suggested from a mathematical model, which reproduces the observed experimental dynamics of E. coli evolution towards virulence. In conclusion, we reveal a molecular mechanism explaining how a specific component of host innate immunity can modulate microbial evolution towards pathogenicity.     

Urea’s Effect on the Ribonuclease A Catalytic Efficiency: A Kinetic, 1H NMR and Molecular Orbital Study

Understanding of protein–urea interactions is one of the greatest challenges to modern structural protein chemistry. Based in enzyme kinetics experiments and ¹H NMR spectroscopic analysis we proposed that urea, at low concentrations, directly interacts with the protonated histidines of the active center of RNase A, following a simple model of competitive inhibition. These results were supported by theoretical analysis based on the frontier molecular orbital theory and suggest that urea might establish a favorable interaction with the cationic amino acids. Our experimental evidence and theoretical analysis indicate that the initials steps of the molecular mechanism of Urea–RNase A interaction passes through the establishment of a three center four electron adduct. Also, our results would explain the observed disruption of the ¹H NMR signals corresponding to H12 and H119 (involved in catalysis) of the RNase A studied in the presence of urea. Our interaction model of urea–amino acids (cationic) can be extended to explain the inactivation of other enzymes with cationic amino acids at the active site.