Combining the Box-Cox power and generalised log transformations to accommodate nonpositive responses in linear and mixed-effects linear models

Transformation of a response variable can greatly expand the class of problems for which the linear regression model or linear mixed-model is appropriate. Beginning with the fundamental work of Box and Cox, maximum-likelihood-like estimation has been applied to select a transformation from among a family of transformations, with the possible goals of achieving approximate normality, removing nonlinearity in a mean function, or stabilizing variance. The Box-Cox power family (BC) of transformations is by far the most common with the Box-Cox methodology, and it requires a strictly positive response. In this article we introduce a new family of transformations that we call the Box-Cox power with nonpositives (BCN) family that allows inclusion of a few nonpositive values. The BCN family is a modification of the basic power family that is inspired by the generalised log, or glog transformation, proposed for use with the more limited goals of stabilizing variance or achieving approximate normality. The glog transformation is itself a special case of the Johnson SU transformation, and we show that the BCN family derived from it is in turn a simple modification of the BC family. Computer code for implementing this family is included in the car package in R (Fox and Weisberg, 2011). The methodology is illustrated using a problem in clinical chemistry. Keywords: Homoscedasticity, Multivariate analysis, Normality

Evaluation of applying statistical process control techniques to daily average feeding behaviors to detect disease in automatically fed group-housed preweaned dairy calves

Group housing and computerized feeding of preweaned dairy calves are gaining in popularity among dairy producers, yet disease detection remains a challenge for this management system. The aim of this study was to investigate the application of statistical process control charting techniques to daily average feeding behavior to predict and detect illness and to describe the diagnostic test characteristics of using this technique to find a sick calf compared with detection by calf personnel. This prospective cross-sectional study was conducted on 10 farms in Minnesota (n = 4) and Virginia (n = 6) utilizing group housing and computerized feeding from February until October 2014. Calves were enrolled upon entrance to the group pen. Calf personnel recorded morbidity and mortality events. Farms were visited either every week (MN) or every other week (VA) to collect calf enrollment data, computer-derived feeding behavior data, and calf personnel-recorded calf morbidity and mortality. Standardized self-starting cumulative sum (CUSUM) charts were generated for each calf for each daily average feeding behavior, including drinking speed (mL/min), milk consumption (L/d), and visits to the feeder without a milk meal (no.). A testing subset of 352 calves (176 treated, 176 healthy) was first used to find CUSUM chart parameters that provided the highest diagnostic test sensitivity and best signal timing, which were then applied to all calves (n = 1,052). Generalized estimating equations were used to estimate the diagnostic test characteristics of a single negative mean CUSUM chart signal to detect a sick calf for a single feeding behavior. Combinations of feeding behavior signals were also explored. Single signals and combinations of signals that included drinking speed provided the most sensitive and timely signal, finding a sick calf up to an average (±SE) of 3.1 ± 8.8 d before calf personnel. However, there was no clear advantage to using CUSUM charting over calf observation for any one feeding behavior or combination of feeding behaviors when predictive values were considered. The results of this study suggest that, for the feeding behaviors monitored, the use of CUSUM control charts does not provide sufficient sensitivity or predictive values to detect a sick calf in a timely manner compared with calf personnel. This approach to examining daily average feeding behaviors cannot take the place of careful daily observation.

Abstract P1-03-06: Improvement in risk predictive value of Nottingham prognostic index by determining GP88 tumor tissue expression for estrogen receptor positive breast cancer patients

Background: The Nottingham Prognostic Index (NPI), which includes nodal status, tumor size and histological grade was established to provide predictive value information on post-surgery survival for primary breast cancer patients. Attempts to improve NPI's performance have included addition of other biomarker expression and morphological features such as vascular invasion. In the present study, we investigated whether expression of the autocrine growth and survival factor GP88 (progranulin), known to be overexpressed in breast cancer, whereas it is negative in normal mammary tissue, would improve NPI's predictive value.

Methods: We examined the tumor tissue GP88 expression by immunohistochemistry (IHC) in formalin fixed paraffin embedded tissue sections from 508 cases of estrogen receptor positive (ER+) invasive ductal carcinoma (IDC) with known clinical outcomes (disease-free and overall survivals) and with known NPI. GP88 IHC tumor tissue expression was determined using an anti-GP88 antibody (clone 6B3) developed in our laboratory. GP88 expression was scored (0, 1+, 2+, 3+) by two board certified pathologists and classified into two IHC score groups of GP88 < 3+ (0, 1+, 2+) and GP88 = 3+. The correlation between GP88 scoring, NPI and disease-free (DFS) and overall survival (OS) outcomes was then examined by Kaplan Meier analysis, Cox proportional Hazard (CPH) ratio and Pearson's C2 test.

Results: Kaplan-Meier survival graphs categorized by NPI scores (< 3.4, 3.4-5.4, and >5.4) and by GP88 expression (< 3+ and 3+) showed that for each NPI subgroup, patients with GP88 IHC score of 3+ had a worse disease-free survival (DFS) and overall survival (OS) than patients within the same NPI subgroup with tumors that had GP88 IHC score < 3+. When adjusted for NPI, high GP88 score was highly significantly associated with recurrence with a hazard ratio of 3.30 (95% CI 2.12 to 5.14).

Conclusions: The data suggest that measuring GP88 tumor tissue expression by IHC at time of diagnosis for breast cancer patients with primary ER+ IDC could provide recurrence prediction and survival information complementary to that provided by the determination of NPI alone and thus may be useful for risk management of patients diagnosed with breast cancer.

Citation Format: Serrero G, Hawkins DM, Bejarano PA, Ioffe O, Tkaczuk KR, Elliott RE, Head JF, Phillips J, Godwin AK, Weaver J, Hicks D, Yue B. Improvement in risk predictive value of Nottingham prognostic index by determining GP88 tumor tissue expression for estrogen receptor positive breast cancer patients [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-03-06.

Abstract P1-02-12: Determination of a serum progranulin (GP88/PGRN) level associated with overall survival in metastatic breast cancer patients

Imaging technologies are the methods of choice in the standard of care (SOC) to monitor therapy response in metastatic breast cancer (MBC) patients. However, such methods are expensive and have limited sensitivity to detect disease response in a timely manner. Measurement of the circulating tumor markers such as CA15-3, CA27.29 and CEA has provided additional minimally invasive methods in disease management of MBC patients. While useful, they have limitations in providing clinicians with a reliable insight into real-time disease monitoring. Understanding of real-time biological processes may provide better biomarkers of the disease state and thus aid real-time clinical management of MBC patients by identifying circulating disease associated biomarkers. Thus, addition of such new circulating biomarkers may improve the management of MBC patients. We have characterized a target biomarker that would fit these criteria, the 88kDa glycoprotein Progranulin (GP88). GP88 is expressed in tumor tissue and not in normal mammary tissue counterpart and is secreted in the circulation of BC patients. Biological studies have established GP88 as one of the critical drivers for breast cancer cell proliferation, survival, invasiveness and drug resistance. Clinical studies have demonstrated that elevated GP88 tumor levels were prognostic for recurrence and that breast cancer patients had a statistically elevated GP88 serum level than healthy individuals. Using tissue and serum tests to detect and quantify GP88 could provide an new strategies for identifying patients at high risk of recurrence and monitoring disease progression in BC patients undergoing therapy. In the present study, we examined whether GP88 serum levels were elevated in MBC patients and whether GP88 serum levels were correlated to patient overall survival.

Under an IRB approved protocol, 92 MBC patients that met the inclusion criteria and were undergoing therapy at the UMGCCC Breast Clinic were consented and enrolled. Clinical and disease characteristics along with serum CA15-3 values were collected as part of the study. Serum samples were collected from each patient during therapy and subsequently the patients were monitored. The serum was stored at -80C until tested for GP88 using a GP88 enzyme linked immunoassay developed in our laboratory.

Statistical analysis using Kaplan-Meier functions established whether there was a correlation between GP88 serum level and overall survival in MBC patients. By analyzing the KM plots at different GP88 cut points, we identified two populations with distinct survival characteristics. When examined more thoroughly the difference in overall survival of patients with <60ng/ml and >60ng/ml was statistically significant (P=0.0002). Correlation analysis of serum GP88 and CA15-3 were performed and will be presented.

We conclude that circulating levels of GP88 in MBC patients are correlated with overall survival. It would appear that patients that can be managed to have a GP88 below 60ng/ml will survive longer. Thus measuring circulating GP88 levels would provide additional information to that available in today's SOC for monitoring. This valuable insight into real-time disease status will assist clinicians in patient management.

Citation Format: Serrero G, Hawkins DM, Yue B, Hicks D, Tait N, Tkaczuk KR. Determination of a serum progranulin (GP88/PGRN) level associated with overall survival in metastatic breast cancer patients [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-02-12.

Fluorescence spectroscopy incorporated in an Optical Biopsy System for the detection of early neoplasia in Barrett's esophagus

Endoscopic surveillance is recommended for patients with Barrett's esophagus (BE) to detect high-grade intraepithelial neoplasia (HGIN) or early cancer (EC). Early neoplasia is difficult to detect with white light endoscopy and random biopsies are associated with sampling error. Fluorescence spectroscopy has been studied to distinguish non-dysplastic Barrett's epithelium (NDBE) from early neoplasia. The Optical Biopsy System (OBS) uses an optical fiber integrated in a regular biopsy forceps. This allows real-time spectroscopy and ensures spot-on correlation between the spectral signature and corresponding physical biopsy. The OBS may provide an easy-to-use endoscopic tool during BE surveillance. We aimed to develop a tissue-differentiating algorithm and correlate the discriminating properties of the OBS with the constructed algorithm to the endoscopist's assessment of the Barrett's esophagus. In BE patients undergoing endoscopy, areas suspicious for neoplasia and endoscopically non-suspicious areas were investigated with the OBS, followed by a correlating physical biopsy with the optical biopsy forceps. Spectra were correlated to histology and an algorithm was constructed to discriminate between HGIN/EC and NDBE using smoothed linear dicriminant analysis. The constructed classifier was internally cross-validated and correlated to the endoscopist's assessment of the BE segment. A total of 47 patients were included (39 males, age 66 years): 35 BE patients were referred with early neoplasia and 12 patients with NDBE. A total of 245 areas were investigated with following histology: 43 HGIN/EC, 66 low-grade intraepithelial neoplasia, 108 NDBE, 28 gastric or squamous mucosa. Areas with low-grade intraepithelial neoplasia and gastric/squamous mucosa were excluded. The area under the receiver operating characteristic curve of the constructed classifier was 0.78. Sensitivity and specificity for the discrimination between NDBE and HGIN/EC of OBS alone were 81% and 58% respectively. When OBS was combined with the endoscopist's assesssment, sensitivity was 91% and specificity 50%. If this protocol would have guided the decision to obtain biopsies, half of the biopsies would have been avoided, yet 4/43 areas containing HGIN/EC (9%) would have been inadvertently classified as unsuspicious. In this study, the OBS was used to construct an algorithm to discriminate neoplastic from non-neoplastic BE. Moreover, the feasibility of OBS with the constructed algorithm as an adjunctive tool to the endoscopist's assessment during endoscopic BE surveillance was demonstrated. These results should be validated in future studies. In addition, other probe-based spectroscopy techniques may be integrated in this optical biopsy forceps system.

Contraception in sea-going service personnel

The right to make an informed choice about contraception should be afforded to every individual serving within the United Kingdom (UK) Armed Forces. This article looks at the responsibilities and approach that healthcare professionals should take within a Primary Care setting, summarises the common contraceptive options available, discusses the associated advantages and disadvantages of each technique, and considers operational factors in a military environment that combine to influence the final contraceptive choice an individual makes. Case Study. A 19-year old Able Rate joined the Royal Navy (RN) and at her joining medical it was noted that she had been on Microgynon™ combined oral contraceptive pill for approximately three years. During this time, her menstrual periods remained light; she never experienced adverse effects, demonstrated good compliance, and was happy to remain on this contraceptive regimen. Over the course of the next eighteen months, she was reviewed by a number of Medical Officers and Civilian Medical Practitioners on a quarterly basis, with Microgynon™ re-prescribed on each occasion. The appropriate Defence Medical Information Capability Programme (DMICP) template was used, with weight, smoking status, compliance and any issues or comments documented accordingly. In December 2010, a discussion regarding long-acting reversible contraception (LARC) was documented for the first time. The patient agreed to give LARC some thought and a review appointment was made for one month. She was subsequently started on the progestogen-only pill Cerazette™. It was noted by the consulting doctor that both the patient's mother and grandmother had a positive history of cerebrovascular events and the combined oral contraceptive pill was discontinued. Upon review at two months, the patient reported that she was content on Cerazette™ and wished to continue with this medication. She was amenorrhoeic, highly compliant, had given up smoking and her weight and blood pressure were stable. However, due to supply issues, it was explained that Cerazette™ was no longer a viable option for her. She had no plans to start a family, and was keen to investigate other contraceptive options. Furthermore, she expressed a particular desire to remain amenorrhoeic, as she was due to deploy overseas in the coming months, and not only wanted to avoid the inconvenience of having her period, but also felt it preferable not to have to take a daily pill when considering the constantly changing time zones. She subsequently had the etonogestrel-releasing subdermal implant Nexplanon™ fitted without complication. She has remained amenorrhoeic throughout and this form of long-acting reversible contraception has particularly suited her busy working role and active lifestyle.

Effects of age and serum 25-OH-vitamin D on serum parathyroid hormone levels

CONTEXT

Several studies define optimal serum 25-hydroxyvitamin D (25-OHD) levels based on serum PTH level reaching an asymptote. However, results differ widely, ranging from 25-OHD levels of 12-44 ng/ml: many studies are constrained by small sample size.

OBJECTIVE

The objective of the study was to determine the relationship between serum PTH and 25-OHD levels and age in a very large reference laboratory database.

DESIGN

This was a detailed cross-sectional analysis of 312,962 paired serum PTH and 25-OHD levels measured from July 2010 to June 2011.

RESULTS

Median PTH levels and the proportion of patients (PTH > 65 pg/ml), from 63 successive 25-OHD frequency classes of 5000 patients, provide smooth, exceptionally well-fitted curves (R(2) = 0.994 and R(2) = 0.995, respectively) without discernible inflection points or asymptotes but with striking age dependencies. Serum 25-OHD was below the recent Institute of Medicine sufficiency guidance of 20 ng/ml in 27% (85,000) of the subjects. More importantly, 40 and 51% of subjects (serum 25-OHD <20 and 10 ng/ml, respectively) had biochemical hyperparathyroidism (PTH > 65 pg/ml).

CONCLUSIONS

This analysis, despite inevitable inherent limitations, introduces several clinical implications. First, median 25-OHD-dependent PTH levels revealed no threshold above which increasing 25-OHD fails to further suppress PTH. Second, the large number of subjects with 25-OHD deficiency and hyperparathyroidism reinforces the Third International Workshop on Asymptomatic Primary Hyper parathyroidism's recommendations to test for, and replete, vitamin D depletion before considering parathyroidectomy. Third, strong age dependency of the PTH-25-OHD relationship likely reflects the composite effects of age-related decline in calcium absorption and renal function. Finally, this unselected large population database study could guide clinical management of patients based on an age-dependent, PTH-25-OHD continuum.

P2-12-32: Association between Progranulin (GP88) Expression and Recurrence Risk for Breast Cancer Patients with Estrogen Receptor Positive Invasive Ductal Carcinoma

Purpose: GP88 (progranulin) is a critical player of breast tumorigenesis for estrogen receptor positive (ER+) breast cancer. Pathological studies showed that GP88 was expressed in invasive ductal carcinoma (IDC), but not in normal mammary tissue, benign lesions or lobular carcinoma. The present study examines GP88 prognostic significance in association with recurrence risk for patients with ER+ IDC.

Patients and Methods: Two retrospective multi-site clinical studies examined GP88 expression by immunohistochemistry (IHC) analysis in paraffin-embedded tumor tissues in correlation with patients’ survival outcomes. The training study established a GP88 cut-off value associated with decreased disease-free (DFS) and overall (OS) survivals. The validation study verified the GP88 cut-off value and compared GP88 prognostic information with other prognostic factors in multivariate analysis.

Results: GP88 expression is associated with a statistically significant increase in recurrence risk for ER+ IDC patients. The training study established that GP88 3+ score by IHC analysis was associated with decreased DFS (p=0.0004) and OS (p=0.0036). The independent validation study verified that GP88 3+ score for the high risk group and demonstrated that GP88 3+ score was associated with a 5.9-fold higher hazard of disease recurrence and a 2.5-fold higher mortality hazard compared to patients with tumor GP88&lt;3+. GP88 remained an independent risk predictor after considering age, nodal status, tumor size, tumor grade, progesterone receptor expression, treatment and disease stage.

Conclusion: Our training and validation studies demonstrate that the survival factor GP88 is a prognostic biomarker, predictive of recurrence risk and increased mortality for ER+ IDC patients, independent from other prognostic factors. These results provide support for measuring GP88 tissue expression for newly diagnosed early stage breast cancer patients.

This work was supported by grants R43CA124179, and U01CA113916 from the National Cancer Institute, grants 07-2007-064 and 02-2010-010 from the Avon Foundation for Women.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-12-32.

Quantitative structure-activity relationship studies of antimalarial compounds from their calculated mathematical descriptors

A wide range of mathematical descriptors that can be calculated without the use of any other experimental data except molecular structure were used to develop models to predict binary (+/-) antimalarial activity of a set of 86 4(1H)-quinolones in two strains of parasite: D6 and TM90-C2B (chloroquine and atovaquone susceptible). The quantitative structure-activity relationship for each strain was of high quality and showed good ability in predicting activity versus inactivity when applied to a data set containing well-known antimalarial drugs.

QSPR checking and validation: a case study with hydroxy radical reaction rate constant

Traditionally, QSAR and QSPR models have been fitted by splitting the available compounds into separate learning and validation sets. The model is then fitted to the learning set and assessed using the validation set. Cross-validation (CV) uses all available compounds for both purposes, so that the full body of available information is brought to bear on both the learning and the validation portions of the study. The price paid for this additional information is a substantially greater computational load. A common mistake in using CV is to omit some of the repetitive computations. This mistake leads to substantial bias in the assessment. A hydroxyl radical reaction rate dataset is used to illustrate the superiority of CV and the pitfalls from its improper execution when modeling using nearest neighbors, paralleling behavior in the well-studied linear model setting.

Bulk tank somatic cell counts analyzed by statistical process control tools to identify and monitor subclinical mastitis incidence

The objective of this study was to examine the relationship between monthly Dairy Herd Improvement (DHI) subclinical mastitis and new infection rate estimates and daily bulk tank somatic cell count (SCC) summarized by statistical process control tools. Dairy Herd Improvement Association test-day subclinical mastitis and new infection rate estimates along with daily or every other day bulk tank SCC data were collected for 12 mo of 2003 from 275 Upper Midwest dairy herds. Herds were divided into 5 herd production categories. A linear score [LNS = ln(BTSCC/100,000)/0.693147 + 3] was calculated for each individual bulk tank SCC. For both the raw SCC and the transformed data, the mean and sigma were calculated using the statistical quality control individual measurement and moving range chart procedure of Statistical Analysis System. One hundred eighty-three herds of the 275 herds from the study data set were then randomly selected and the raw (method 1) and transformed (method 2) bulk tank SCC mean and sigma were used to develop models for predicting subclinical mastitis and new infection rate estimates. Herd production category was also included in all models as 5 dummy variables. Models were validated by calculating estimates of subclinical mastitis and new infection rates for the remaining 92 herds and plotting them against observed values of each of the dependents. Only herd production category and bulk tank SCC mean were significant and remained in the final models. High R2 values (0.83 and 0.81 for methods 1 and 2, respectively) indicated a strong correlation between the bulk tank SCC and herd's subclinical mastitis prevalence. The standard errors of the estimate were 4.02 and 4.28% for methods 1 and 2, respectively, and decreased with increasing herd production. As a case study, Shewhart Individual Measurement Charts were plotted from the bulk tank SCC to identify shifts in mastitis incidence. Four of 5 charts examined signaled a change in bulk tank SCC before the DHI test day identified the change in subclinical mastitis prevalence. It can be concluded that applying statistical process control tools to daily bulk tank SCC can be used to estimate subclinical mastitis prevalence in the herd and observe for change in the subclinical mastitis status. Single DHI test day estimates of new infection rate were insufficient to accurately describe its dynamics.

Quantitative structure-activity relationship modeling of insect juvenile hormone activity of 2,4-dienoates using computed molecular descriptors

Juvenile hormone (JH) activity of one hundred and eighty 2,4-dienoates reported for the larvae/pupae of six insect species was modeled using 915 atom pairs and 258 global molecular descriptors (topological and geometrical). Ridge regression, principal component regression and partial least square regression methods were used to model each of the JH activities. The use of all of the available parameters did not yield any good models, and extensive predictor trimming was necessary to improve the models. Ridge regression was found to give the best results among the three statistical tools used. The top ten molecular descriptors selected based on the t-statistic for each of the six models were found to be mostly atom pairs containing heteroatoms and topochemical descriptors. This suggests the importance of the chemical nature of the ligand rather than mere space-filling as the basis of the JH bioactivity. The residual plots indicate the existence of some non-linear relations, and recursive partitioning was used to capture any nonlinear relation between the bioassays and the molecular descriptors.

Establishing pediatric immune response zones using the Cylex® ImmuKnowTM assay

For all transplant patients, the transplant physician must balance the risk of rejection caused by under-immunosuppression against the risk of drug toxicity, secondary infections and post-transplant lymphoproliferative disorder with over-immunosuppression. A Food and Drug Administration (FDA)-approved in vitro assay, the Cylex ImmuKnow assay, provides a global assessment of cellular immune function to help monitor the immune status of immunosuppressed patients. This assay uses the plant lectin phytohemagglutinin to stimulate lymphocytes; an ATP assay is then used to measure the degree of activation of CD4 T cells. However, the normal values for this assay were developed with healthy adult patients. In this study, we determined the normal ranges for the ImmuKnow assay in healthy children and compared those values to levels obtained in healthy adults and in stable pediatric renal transplant patients. We found that healthy children 12 yr of age and older showed immune function levels indistinguishable from adults, while healthy children under 12 had significantly lower immune function levels than adults. For adults, the ImmuKnow assay zones (in ng/mL ATP) of strong, moderate and low immune function correspond to >525, 225 to 525, and <225. In children under 12, we found the corresponding zones to be >395, 175-395 and <175 ng/mL. The median value for normal adults is 415, whereas it is only 295 for children <12 yr of age and this value decreases to 165 in stable renal transplant patients <12 yr of age (compared with 258 for stable adult renal transplant patients). Thus, this study provides critical information necessary to utilize the ImmuKnow assay with pediatric patients. In adults, the degree of immune function as assessed by the ImmuKnow assay helps to predict patients at risk for infection or rejection. If further studies in pediatric patients document the same and is true for children, then the ImmuKnow assay will provide a useful adjunct tool to prevent over- or under-immunosuppression as newly developed drugs are utilized or drug treatment is altered because of drug side effects, toxicity, concurrent illnesses or rejection.

A variance-stabilizing transformation for gene-expression microarray data

MOTIVATION

Standard statistical techniques often assume that data are normally distributed, with constant variance not depending on the mean of the data. Data that violate these assumptions can often be brought in line with the assumptions by application of a transformation. Gene-expression microarray data have a complicated error structure, with a variance that changes with the mean in a non-linear fashion. Log transformations, which are often applied to microarray data, can inflate the variance of observations near background.

RESULTS

We introduce a transformation that stabilizes the variance of microarray data across the full range of expression. Simulation studies also suggest that this transformation approximately symmetrizes microarray data.

Prediction of tissue-air partition coefficients: a comparison of structure-based and property-based methods

Three linear regression methods were used to develop models for the prediction of rat tissue-air partition coefficient (P). In general, ridge regression (RR) was found to be superior to principal component regression (PCR) and partial least squares regression (PLS). A set of 46 diverse low molecular-weight volatile chemicals was used to model fat-air, liver-air and muscle-air partition coefficients for male Fischer 344 rats. Comparisons were made between models developed using descriptors based solely on molecular structure and those developed using experimental properties, including saline-air and olive oil-air partition coefficients, as independent variables, indicating that the structure-property correlations are comparable to the property-property correlations. Multiple structure-based models were developed utilizing various classes of structural descriptors based on level of complexity, i.e. topostructural (TS), topochemical (TC), 3-dimensional (3D) and calculated octanol-water partition coefficient. In most cases, the structure-based models developed using only the TC descriptors were found to be superior to those developed using other structural descriptor classes. Haloalkane subgroups were modeled separately for comparative purposes, and although models based on the congeneric compounds were superior, the models developed on the complete sets of diverse compounds were acceptable. Comparisons were also made with respect to the types of descriptors important for partitioning across the various media.

Some issues in resolution of diagnostic tests using an imperfect gold standard

As a subject's true disease status is seldom known with certainty, it is necessary to compare the performance of new diagnostic tests with those of a currently accepted but imperfect 'gold standard'. Errors made by the gold standard mean that the sensitivity and specificity calculated for the new test are biased, and do not correctly estimate the new method's sensitivity and specificity. The traditional approach to this problem was 'discrepant resolution', in which the subjects for whom the two methods disagreed were subjected to a third 'resolver' test. Recent work has pointed out that this does not automatically solve the problem. A sounder approach goes beyond the discordant test results and tests at least some of the subjects with concordant results with the resolver also. This leaves some issues unresolved. One is the basic question of the direction of biases in various estimators. We point out that this question does not have a simple universal answer. Another issue, if one is to test a sample of the subjects with concordant results rather than all cases, is how to compute estimates and standard errors of the measures of test performance, notably sensitivity and specificity of the test method relative to the resolver. Expressions for these standard errors are given and illustrated with a numeric example. It is shown that using just a sample of subjects with concordant results may lead to great savings in assays. The design issue of how many concordant cells to test depends on the numbers of concordants and discordants. The formulae given show how to evaluate impact of different choices for these numbers and hence settle on a design that gives the required precision of estimates.

QSAR with few compounds and many features

Fitting quantitative structure-activity relationships (QSAR) requires different statistical methodologies and, to some degree, philosophies depending on the "shape" of the data matrix. When few features are used and there are many compounds, it is a reasonable expectation that good feature subset selection may be made and that nonlinearities and nonadditivities can be detected and diagnosed. Where there are many features and few compounds, this is unrealistic. Methods such as ridge regression RR, PLS, and principal component regression PCR, which abjure feature selection and rely on linearity may provide good predictions and fair understanding. We report a development of ridge regression for the underdetermined case by using generalized cross-validation to choose the ridge constant and perform F-tests for additional information. Conventional regression diagnostics can be used in followup to identify nonlinearities and other departures from model. We illustrate the approach with QSAR models of four data sets using calculated molecular descriptors.

Validity of an estimator of reproductive success

Lifetime reproductive success is a major component of individual fitness and a central dependent variable for the study of natural selection. For long-lived animals, such as apes or baboons, assessment of lifetime reproductive success requires observations of identified individuals in continuous, long-term studies from which it is difficult and often impossible to obtain an adequate sample of necessary reproductive and survival data. This situation can be alleviated by the availability of a valid measure that uses incomplete reproductive histories to estimate the lifetime reproductive success of individuals. The validity of one such estimator was tested by determining if, after 10.5 years of studying free-ranging female baboons, it predicted lifetime reproductive success obtained from full reproductive histories after 21.5 years. Validity was evaluated for seven criteria of success, ranging from the number of a female's live births to the number of her offspring that reached the age of 72 months. Moderate to good prediction of lifetime reproductive success by the estimator was found for criteria of offspring living to 36 months or more. After 10.5 years, complete reproductive life spans were available for only eight females. Using the estimator, analytic potential, via sample representativeness and size, was improved at 10.5 years by an increase from a sample of eight to between 34 and 62, depending upon the criterion used, and at 21.5 years from 39 to 70. With a valid estimator, the opportunity to study lifetime reproductive success of a long-lived species is substantially improved without having to depend upon rarely available, uninterrupted data collection for 20-60 years.

An invitation to join in difficulty: realizing the deeper promise of group psychotherapy

Group psychotherapy has proven effective in treating an impressive array of specific problems. Clinical experience has also shown its utility in alleviating the more general, and very costly, alienation and pain with which the human condition is often fraught. Recent changes in health care delivery reflect a marketplace that emphasizes cost containment, brevity and specificity of treatment, and a narrowed focus of training. Techniques are often valued above relationships. In this climate, we are in danger of losing sight of group therapy's deeper promise; the experiential lesson that human beings must move toward each other with involvement and a commitment to understanding in response to the inevitable difficulties that arise as we share time, space, and resources.

Analysis of a 2(9) full factorial chemical library

Robotic synthesis is making possible the synthesis of large, systematically designed sets of compounds. We analyze a 512-compound set that is a 2(9) full factorial experimental design using a recursive partitioning algorithm, FIRM, and a high-dimension visualization tool, TempleMVV. These techniques are used to quickly and easily identify the main trends in the data set and also identify unusual observations. We show that analytical and visualization methods can be used synergistically to analyze a large, complex, high-dimensional data set. We also show that a fractional factorial design of 128 compounds would give essentially the same information.

Applications of geostatistics in plant nematology

The application of geostatistics to plant nematology was made by evaluating soil and nematode data acquired from 200 soil samples collected from the A(p) horizon of a reed canary-grass field in northern Minnesota. Geostatistical concepts relevant to nematology include semi-variogram modelling, kriging, and change of support calculations. Soil and nematode data generally followed a spherical semi-variogram model, with little random variability associated with soil data and large inherent variability for nematode data. Block kriging of soil and nematode data provided useful contour maps of the data. Change of snpport calculations indicated that most of the random variation in nematode data was due to short-range spatial variability in the nematode population densities.

Correlation of Edaphic Factors with Plant-parasitic Nematode Population Densities in a Forage Field

Two hundred soil samples from the A(p) horizon of a reed canarygrass field overlaying several different but related soils in northern Minnesota were analyzed for plant-parasitic nematodes and 22 edaphic factors. Pratylenchus penetrans was the predominant nematode taxon. Others were Aglenchus agricola, Tylenchorhynchus spp., Heterodera trifolii, Paratylenchus spp., Tylenchus maius, and Criconemella sp. Five nematode taxa, P. penetrans, A, agricola, Tylenchorhynchus spp., H. trifolii, and Paratylenchus spp., were correlated with particle size, Tylenchus maius and Criconemella sp. were correlated with effective cation exchange capacity. Nematode field spatial arrangements were related to a combination of statistically significant positive and negative soil factor effects on the nematode populations. Contour maps derived by geostatistical techniques were used to visually validate statistically significant correlations of nematode and soil data. Contour mapping to supplement traditional statistical techniques can be used to achieve a more holistic approach to studies of nematode-soil interrelationships.

Cryopreservation of semen from unique lines of chicken germ plasm

Frozen semen is a practical means of preserving valuable germ plasm. Monitored samples of semen cryopreserved with glycerol for heterozygous, dominant marker stocks and for nine chromosomal rearrangement lines had sufficiently high fertility for germ-line retrieval. The results also indicated a potential for the genetic selection of certain lines for the freezability of spermatozoa, since stock and line differences in fertility occurred when previously frozen semen was used for insemination. Freezing the semen of stocks routinely reproduced provides insurance against possible disasters.

Estimation of the relative fertilizing ability of frozen chicken spermatozoa using a heterospermic competition method

Semen was collected from Rhode Island Red and White Leghorn roosters. After adjusting the sperm number from both breeds, half of the semen was frozen-thawed in the presence of glycerol. Frozen and unfrozen semen from both breeds was mixed 1:1 with all four possible combinations and inseminated into Rhode Island Red hens. Feather colour of chicks was used to determine which breed fertilized the eggs. Results showed that sperm cells retained 19.7% [95% confidence interval = (12.8, 30.4)] of the relative fertilizing ability after freezing. Furthermore, Rhode Island Red spermatozoa had 1.5 times [95% confidence interval = (1.1, 2.0)] higher relative fertilizing ability than did White Leghorn spermatozoa. The heterospermic competition assay method is a powerful tool for estimating the relative fertilizing ability of the sperm cells.

Reactive affective cycles and lithium: a caveat

The author has been referred increasing numbers of patients who were inappropriately treated with Lithium for cyclic "depressions" which were in fact reactive-depressive episodes resulting from cyclic patterns of interpersonal behavior. This paper gives two case reports to demonstrate the phenomenon of reactive-affective cycles, their clinical presentations and the problems arising from treatment with medications. The structure of cycles of interpersonal behavior and their presentation in therapy is described and treatment strategies are suggested.

Non-ejection systolic clicks and mitral systolic murmurs in black schoolchildren of Soweto, Johannesburg

A survey was conducted on 12 050 Black schoolchildren, aged 2 to 18 years, in the South Western Townships of Johannesburg (Soweto), and the prevalence of non-ejection systolic clicks and late systolic murmurs was determined. One or both of these auscultatory findings were detected in 168 children, yielding a prevalence rate of 13-99 per 1000 in the school population. A female preponderance of 1-9:1 was present and there was a strong linear increase in prevalence with age, with a peak rate of 29-41 per 1000 in 17-year-old children. A non-ejection click was the only abnormal auscultatory finding in 123 children (73%) and a mitral systolic murmur in 8 (5%), whereas in 37 (22%) both these findings were present. Of the latter 37 children, the murmur was late systolic in 32; in 5 it was early systolic. Auscultation in different postures was important in the detection of both non-ejection clicks and mitral systolic murmurs. Experience in the detection of these auscultatory findings influenced the frequency with which they were heard. Electrocardiographic abnormalities compatible with those previously described in the billowing mitral leaflet syndrome were present in 11 of 158 children. The aetiology of these auscultatory findings in this community remains unknown. In the same survey, a high prevalence rate of rheumatic heart disease was recorded and the epidemiology of the non-ejection clicks and these mitral systolic murmurs showed similarties to that of rheumatic heart disease. Though the specific billowing mitral leaflet syndrome almost certainly accounts for some of these auscultatory findings, a significant proportion may have early rheumatic heart disease. Further elucidation of this problem is necessary.

The prediction of lymph node metastases from oral squamous carcinoma

In an attempt to predict which cases of oral and oropharyngeal squamous carcinoma are likely to metastasize to regional lymph nodes a series of 898 cases was grouped according to site, size, grade of histological differentiation, and presence or absence of histologically confirmed regional lymph node metastases. The results were analysed by a logistic multiple regression analysis. They showed that the sites may be divided into three clusters, Cluster 1 consists of tumors of lip, floor of mouth, cheek mucosa, hard palate, and gingiva. These are not significantly different as regards metastasis rate. Cluster 2 consists of tumors of the anterior two-thirds of tongue and has a higher tendency to metastasis than those in Cluster 1. Lesions of the posterior third of tongue and oropharynx form Cluster 3 which exhibits the greatest tendency to metastasis. Sizes of primary lesions are clustered in groups of lesions less than 3 cm, those 3 to less than 4 cm, and those 4 cm or larger, in ascending tendency to metastasis. Well-differentiated and moderately differentiated tumors are not significantly different in their tendency to metastasize and may be reduced to a single cluster, whereas poorly differentiated tumors have a markedly higher metastasis rate. Using these clusters it has been possible to predict the logistically transformed probability of metastasis to a high degree of accuracy (R=0.9398). From this we conclude that if for a given tumor we know to which site, size or differentiation cluster it belongs, we can then estimate its probability of metastasising.

Epidemiology of rheumatic heart disease in black shcoolchildren of Soweto, Johannesburg

A survey to determine the prevalence of rheumatic heart disease (R.H.D.) in Black children was conducted in the creeches and primary schools of the South Western Townships of Johannesburg (Soweto). A total of 12 050 Black children were examined by 10 cardiologists in May to October 1972. The overal prevalence rate of R.H.D. was 6.9 per 1000, with a peak rate of 19.2 per 1000 in children of the seventh school grade. The maximal age incidence was 15-18 years and there was a female preponderance of 1 6:1. A rise in prevalence occurred with increasing family size. Most children (92%) were asymptomatic, and in 82.5% R.H.D. was diagnosed for the first time during the school survey. The commonest valve lesion was mitral regurgitation, which was present in 93% and occurred as an isolated lesion in 47.5%. Lancefield's group A beta-haemolytic streptococcus was isolated from the throats of 52 per 1000 Soweto children. The auscultatory features of a non-ejection systolic click and late systolic murmur were prevalent (13.9 per 1000) and had several epidemiological factors in common with R.H.D. A comprehensive preventative campaign is urgently needed in South Africa, directed at both primary and secondary prophylaxis of R.H.D. The socioeconomic status of the community must be improved if optimal prevention is to be achieved.

Rabies in New Mexico cavern bats

Images