The Effect of Target Vector Selection on the Invariance of Classifier Performance Measures

In this paper, the multiclass supervised training problem is considered when a discrete set of classes is assumed. Upon generating affine models for finite data sets, we have observed the invariance of certain measures of performance after a trained classifier has been presented with test data of unknown classification. Specifically, after constructing mappings between training vectors and their desired targets, the class membership and ranking of test data has been found to remain either invariant or close to invariant under a transformation of the set of target vectors. Therefore, we derive conditions explaining how this type of invariance can arise when the multiclass problem is phrased in the context of linear networks. A bioinformatics example is then presented in order to demonstrate various principles outlined in this work.     

Strategic questions for consumer-based health communications

Using the consumer-oriented approach of social and commercial marketers, this article presents a process for crafting messages designed to improve people's health behaviors. The process, termed consumer-based health communications (CHC), transforms scientific recommendations into message strategies that are relevant to the consumer. The core of CHC is consumer research conducted to understand the consumer's reality, and thereby allowing six strategic questions to be answered. The immediate result of the CHC process is a strategy statement--a few pages that lay out who the target consumer is, what action should be taken, what to promise and how to make the promise credible, how and when to reach him or her, and what image to convey. The strategy statement then guides the execution of all communication efforts, be they public relations, mass media, direct marketing, media advocacy, or interpersonal influence. It identifies the most important "levers" for contact with the consumer. Everyone from creative specialists through management and program personnel can use the strategy statement as a touchstone to guide and judge the effectiveness of their efforts. The article provides a step by step illustration of the CHC process using the 5 A Day campaign as an example.     

Approximate regular expression pattern matching with concave gap penalties

Given a sequenceA of lengthM and a regular expressionR of lengthP, an approximate regular expression pattern-matching algorithm computes the score of the optimal alignment betweenA and one of the sequencesB exactly matched byR. An alignment between sequencesA=a₁a₂ ... a_M andB=b₁b₂... b_N is a list of ordered pairs, (i₁,j₁), (i₂j₂), ..., (i_t,j_tt) such that i_k < i_k+1 and j_k < j_k+1. In this case the alignmentaligns symbols a_ik and b_jk, and leaves blocks of unaligned symbols, orgaps, between them. A scoring schemeS associates costs for each aligned symbol pair and each gap. The alignment's score is the sum of the associated costs, and an optimal alignment is one of minimal score. There are a variety of schemes for scoring alignments. In a concave gap penalty scoring schemeS={, w}, a function (a, b) gives the score of each aligned pair of symbolsa andb, and aconcave function w(k) gives the score of a gap of lengthk. A function w is concave if and only if it has the property that, for allk > 1, w(k + 1) –w(k) w(k) –w(k –1). In this paper we present an O(MP(logM + log² P)) algorithm for approximate regular expression matching for an arbitrary and any concavew. This work was supported in part by the National Institute of Health under Grant RO1 LM04960.     

An unusual case of Creutzfeldt-Jakob disease

A patient with histopathologically verified sporadic Creutzfeldt-Jakob disease (CJD) presented initially with diplopia, sleep disturbances, and L-dopa-responsive parkinsonism. After more than a year of slow progression, he did not become demented, and failed to fulfill the clinical criteria for possible CJD. No clinical examinations currently proposed to detect CJD showed the disease. CJD should be in the differential diagnosis of "parkinson plus" syndromes until a different etiology has been found or a histopathologic examination performed.     

NK cells can recognize different forms of class I MHC

NK recognition and lysis of targets are mediated by activation receptor(s) whose effects may be over-ridden by inhibitory receptors recognizing class I MHC on the target. Incubation of normal lymphoblasts with a peptide that can bind to their class I MHC renders them sensitive to lysis by syngeneic NK cells. By binding to class I MHC, the peptide alters or masks the target structure recognized by an inhibitory NK receptor(s). This target structure is most likely an "empty" dimer of class I heavy chain and beta2m as opposed to a "full" class I trimer formed by binding of specific peptide that is recognized by CTL.     

Rupture of a hepatic artery aneurysm and renal infarction: 2 complications of fibromuscular dysplasia that mimic vasculitis

Fibromuscular dysplasia (FMD) is one of the most important mimics of vasculitis. FMD is more prevalent in women and may cause infarcts of organ systems including the brain, intestines, and extremities. These acute clinical presentations often raise the suspicion of vasculitis and therefore rheumatologists will be asked to evaluate these patients. One of the frequent manifestations of FMD is renal artery involvement. Angiographic evaluation often shows a "string of beads" appearance of the renal arteries but sometimes only tapering of the vascular lumen or cutoffs may be seen. In contrast, hepatic artery involvement with FMD occurs much less frequently and may initially present with life threatening intraabdominal bleeding. Celiac angiograms of these patients often show aneurysms, a presentation similar to vasculitis. Only a high level of suspicion for FMD and careful evaluation of clinical facts will lead to correct diagnosis. This will prevent the wrong diagnosis of vasculitis and subsequent treatment with potentially toxic medications. We present 2 cases of FMD involving the renal and hepatic arteries that illustrate this diagnostic challenge.     

Multiple evolutionary origins of the fungus causing Panama disease of banana: concordant evidence from nuclear and mitochondrial gene genealogies

Panama disease of banana, caused by the fungus Fusarium oxysporum f. sp. cubense, is a serious constraint both to the commercial production of banana and cultivation for subsistence agriculture. Previous work has indicated that F. oxysporum f. sp. cubense consists of several clonal lineages that may be genetically distant. In this study we tested whether lineages of the Panama disease pathogen have a monophyletic origin by comparing DNA sequences of nuclear and mitochondrial genes. DNA sequences were obtained for translation elongation factor 1alpha and the mitochondrial small subunit ribosomal RNA genes for F. oxysporum strains from banana, pathogenic strains from other hosts and putatively nonpathogenic isolates of F. oxysporum. Cladograms for the two genes were highly concordant and a partition-homogeneity test indicated the two datasets could be combined. The tree inferred from the combined dataset resolved five lineages corresponding to "F. oxysporum f. sp. cubense" with a large dichotomy between two taxa represented by strains most commonly isolated from bananas with Panama disease. The results also demonstrate that the latter two taxa have significantly different chromosome numbers. F. oxysporum isolates collected as nonpathogenic or pathogenic to other hosts that have very similar or identical elongation factor 1alpha and mitochondrial small subunit genotypes as banana pathogens were shown to cause little or no disease on banana. Taken together, these results indicate Panama disease of banana is caused by fungi with independent evolutionary origins.     

Cancer pain management in developing countries: a mosaic of complex issues resulting in inadequate analgesia

OBJECTIVE: To determine the prevalence of suicidal ideation and the symptomatic profile as well as to identify the sociodemographic characteristics highlighting female adolescents with the highest scores on suicidal ideation in adolescents living in Mexico City: students and suicidal patients. MATERIAL AND METHODS: A cross-sectional and ex-post-facto study was carried out in order to analyze information from two samples: 1,712 junior and junior high school women students (representative student sample in Mexico City), and 30 adolescents inpatient hospitalized for her suicide attempts (clinical sample). RESULTS: Prevalence of presence as well as persistence of suicidal ideation were higher in the clinical sample, nevertheless 11.8% of the school sample had everyone of the symptoms in a range of 1 to 7 days. The most persistent of the ideation symptoms was: "My family would be better if I were dead"; and in the clinical sample it was "I thought about killing myself". Finally, the sociodemographic characteristics that best matched the student girls having the highest scores in suicidal ideation were: to be on junior high school, to get low grades, to acknowledge school performance as bad and to have interrupted her studies. The characteristics that highlight the girls with a first attempt were analyzed in the clinic sample in comparison with those ones with two or more attempts. It was significant that girls in the second group were living only with one parent and they thought in the last attempt that their dead would be possible or certain. CONCLUSIONS: The prevalence of suicidal ideation was important in the school sample. If it is considered that this psychological construct has a strong association with suicidal attempt, and multiple suicide conduct, then is a priority to detect adolescents in risk and to make preventive efforts, considering the sociodemographic characteristics that configurate in risk for suicidal ideation.     

Crystal growth in dental enamel: the role of amelogenins and albumin

Amelogenin-mineral interactions were investigated using an in vitro binding approach. Rat incisor enamel matrix proteins (mainly amelogenins) were dissolved in synthetic enamel fluid and allowed to equilibrate with deproteinised developing enamel crystals. The results showed that amlogenin proteins of 21, 23, 24, 26 and 27-kDa (corresponding to nascent and partially degraded amelogenins) were associated with the crystals whilst the lower Mr amelogenins (< 21 KDa) remained free in the synthetic enamel fluid. These data suggest the nascent and partially degraded amelogenins may interact with developing enamel crystals and could influence their growth. Albumin-mineral interactions were investigated by extracting developing rat incisor enamel with synthetic enamel fluid. Insoluble material (including the enamel crystals) was then further extracted with 0.1 M phosphate buffer (pH 7.4) to desorb any mineral bound proteins. Western blotting using anti-albumin antibodies showed that almost all of the albumin from the secretory stage enamel and a significant proportion of the albumin present in early transition stage was extractable in the synthetic enamel fluid. However, synthetic enamel fluid did not extract albumin from late transition or maturation stage tissue, which could only be removed following further extraction with phosphate buffer. Albumin degradation was apparent during the transition and maturation stages, where it is degraded and ultimately removed. This binding pattern may be related to amelogenin degradation and removal during the transition stage, permitting albumin access to the previously obscured crystal surfaces. That the secretory stage matrix appears to "protect" secretory stage crystals from albumin may be an important consideration in the aetiology of enamel hypoplasias (i.e. incomplete crystal growth) and when using dissociative extraction procedures for the identification of mineral bound proteins.     

Assessment of the nutritional state of dialysis patients

Obstructive sleep apnea syndrome (OSAS) has been associated with a higher than normal cardiovascular morbidity and mortality. Some OSAS patients lack the sleep-related, nocturnal decrease, or "dip," in blood pressure which is seen in normal individuals. These subjects, called "non-dippers," may be at greater risk for cardiovascular problems. We studied 40 OSAS patients (including 3 women) and 6 control subjects, all identified by polysomnography, for nocturnal blood pressure "dipping." We performed a second nocturnal polysomnogram to determine their apnea and hypopnea indices, (A + H)I, and oxygen saturation levels at the beginning of the study and then initiated 48 hours of ambulatory blood pressure monitoring, with data points collected every 30 minutes. Controls, which included one hypertensive subject, were all dippers. Nineteen OSAS subjects (48% of OSAS individuals) were systolic non-dippers and only 9 of them (22.5%) were diastolic non-dippers. We considered the following clinical variables as potential predictors of non-dipping: age, body mass index, respiratory disturbance index, years of reported loud snoring by bed partners, lowest oxygen saturation during nocturnal sleep, and percentage of sleep time spent with oxygen saturation below 90%. Multiple regression analyses indicated respiratory disturbance index as the only significant variable for systolic (p = 0.04) and diastolic (p = 0.03) blood pressure non-dipping. When we forced the following two nonsignificant variables into the model, they showed a very meager impact: number of years with reported loud snoring (p = 0.4 and p = 0.5, respectively for systolic and diastolic blood pressure non-dipping) and age (p = 0.5 and p = 0.6). The calculated model explained only a low percentage of the variance with an r2 of 0.25 and 0.26 for systolic and diastolic blood pressure non-dipping, respectively. Analysis of hypertension/normotension and dipping/non-dipping failed to show a significant relationship in the studied population. Fifty percent of the normotensive OSAS subjects were non-dippers and 43% of the hypertensive OSAS subjects were also non-dippers. We found a relationship between increasing respiratory disturbance index and increasing average 24-hour systolic blood pressure only when OSAS subjects were non-dippers and hypertensive.