John T. Casagrande

Large-scale gene function analysis with the PANTHER classification system

The PANTHER (protein annotation through evolutionary relationship) classification system (http://www.pantherdb.org/) is a comprehensive system that combines gene function, ontology, pathways and statistical analysis tools that enable biologists to analyze large-scale, genome-wide data from sequencing, proteomics or gene expression experiments. The system is built with 82 complete genomes organized into gene families and subfamilies, and their evolutionary relationships are captured in phylogenetic trees, multiple sequence alignments and statistical models (hidden Markov models or HMMs). Genes are classified according to their function in several different ways: families and subfamilies are annotated with ontology terms (Gene Ontology (GO) and PANTHER protein class), and sequences are assigned to PANTHER pathways. The PANTHER website includes a suite of tools that enable users to browse and query gene functions, and to analyze large-scale experimental data with a number of statistical tests. It is widely used by bench scientists, bioinformaticians, computer scientists and systems biologists. In the 2013 release of PANTHER (v.8.0), in addition to an update of the data content, we redesigned the website interface to improve both user experience and the systems analytical capability. This protocol provides a detailed description of how to analyze genome-wide experimental data with the PANTHER classification system.

PANTHER version 10: expanded protein families and functions, and analysis tools

PANTHER (Protein Analysis THrough Evolutionary Relationships, http://pantherdb.org) is a widely used online resource for comprehensive protein evolutionary and functional classification, and includes tools for large-scale biological data analysis. Recent development has been focused in three main areas: genome coverage, functional information (‘annotation’) coverage and accuracy, and improved genomic data analysis tools. The latest version of PANTHER, 10.0, includes almost 5000 new protein families (for a total of over 12 000 families), each with a reference phylogenetic tree including protein-coding genes from 104 fully sequenced genomes spanning all kingdoms of life. Phylogenetic trees now include inference of horizontal transfer events in addition to speciation and gene duplication events. Functional annotations are regularly updated using the models generated by the Gene Ontology Phylogenetic Annotation Project. For the data analysis tools, PANTHER has expanded the number of different ‘functional annotation sets’ available for functional enrichment testing, allowing analyses to access all Gene Ontology annotations—updated monthly from the Gene Ontology database—in addition to the annotations that have been inferred through evolutionary relationships. The Prowler (data browser) has been updated to enable users to more efficiently browse the entire database, and to create custom gene lists using the multiple axes of classification in PANTHER.

Oral contraceptive use and early abortion as risk factors for breast cancer in young women.

A case-control study was conducted in Los Angeles County, California, of 163 very young breast-cancer cases (all aged 32 or less at diagnosis) to investigate the role, if any, of oral contraceptives (OC) in the development of the disease. OC use before first full-term pregnancy (FFTP) was associated with an elevated risk, which increased with duration of OC use (relative risk approximately 2.2 at 6 years of use, P < 0.01). This increased risk could not be explained by other risk factors. OC use after FFTP was not associated with any change in risk. A first-trimester abortion before FFTP, whether spontaneous or induced, was associated with a 2.4-fold increase in breast-cancer risk (P < 0.005).

The Relationship Between Breast Cancer and Augmentation Mammaplasty: An Epidemiologic Study

Surgical implantation of breast prostheses for cosmetic purposes has become increasingly popular, and by 1981, it was estimated that three-quarters of a million women had had such an operation. The long-term potential risks, particularly of breast cancer, of such procedures have not been properly investigated. To evaluate the potential breast cancer risk, we have conducted a retrospective cohort study of 3111 women followed through various public and medical records for a total of 18,476 person-years, with a median of 6.2 years per person. The cases of breast cancer were detected by means of a computerized match with the Los Angeles County Cancer Surveillance Program, a population-based cancer registry. Overall, 15.7 breast cancer cases were expected and 9 were observed, a nonsignificant deficit [standardized incidence ratio (SIR) = 57 percent, 95 percent confidence limits: 26 percent, 109 percent]. The cancers were generally diagnosed at an early stage. Among the 573 women aged 40 or older at implantation, 7.1 cases were expected and 8 were observed (SIR = 113 percent). In women whose implants were performed before the age of 40, only 1 case was observed whereas 8.6 cases were expected (SIR = 12 percent, 95 percent confidence limits: 0.3 percent, 65 percent), a significant difference. These data do not support an increased risk of breast cancer following augmentation mammaplasty. The low breast cancer rate in women having augmentation mammaplasty at a young age suggests that many such women may have a reduced amount of breast tissue, but data on this are unavailable.

A new procedure for the statistical evaluation of intrauterine contraception

An improved statistical method for assessing the efficacy of intrauterine contraception is presented. This method, called the log-rank method, is based on a daily life table, which yields the best estimates of the cumulative net termination probabilities. It provides a sensitive chi-square test statistic to detect differences among these probabilities for two or more contraceptive devices. When three or more devices are compared, the chi-square statistic may be partitioned into additive components representing comparisons among groups of devices. The log-rank method also permits adjusting for differences among subgroups by controlling for factors such as parity or age. Although calculations may be performed on a desk calculator, a FORTRAN-IV computer program is made available to interested researchers.

Industrial Air Pollution: Possible Effect on Lung Cancer

Higher lung cancer mortality rates occurred in males living in certain heavily industrialized areas of Los Angeles County, California. These areas were characterized by elevated concentrations of benzo[a]pyrene and other polynuclear aromatic hydrocarbons of primarily industrial origin in the soil and air.

Anthropometric and Hormonal Risk Factors for Male Breast Cancer: Male Breast Cancer Pooling Project Results

BACKGROUND The etiology of male breast cancer is poorly understood, partly because of its relative rarity. Although genetic factors are involved, less is known regarding the role of anthropometric and hormonally related risk factors. METHODS In the Male Breast Cancer Pooling Project, a consortium of 11 case-control and 10 cohort investigations involving 2405 case patients (n = 1190 from case-control and n = 1215 from cohort studies) and 52013 control subjects, individual participant data were harmonized and pooled. Unconditional logistic regression generated study design-specific (case-control/cohort) odds ratios (ORs) and 95% confidence intervals (CIs), with exposure estimates combined using fixed effects meta-analysis. All statistical tests were two-sided. RESULTS Risk was statistically significantly associated with weight (highest/lowest tertile: OR = 1.36; 95% CI = 1.18 to 1.57), height (OR = 1.18; 95% CI = 1.01 to 1.38), and body mass index (BMI; OR = 1.30; 95% CI = 1.12 to 1.51), with evidence that recent rather than distant BMI was the strongest predictor. Klinefelter syndrome (OR = 24.7; 95% CI = 8.94 to 68.4) and gynecomastia (OR = 9.78; 95% CI = 7.52 to 12.7) were also statistically significantly associated with risk, relations that were independent of BMI. Diabetes also emerged as an independent risk factor (OR = 1.19; 95% CI = 1.04 to 1.37). There were also suggestive relations with cryptorchidism (OR = 2.18; 95% CI = 0.96 to 4.94) and orchitis (OR = 1.43; 95% CI = 1.02 to 1.99). Although age at onset of puberty and histories of infertility were unrelated to risk, never having had children was statistically significantly related (OR = 1.29; 95% CI = 1.01 to 1.66). Among individuals diagnosed at older ages, a history of fractures was statistically significantly related (OR = 1.41; 95% CI = 1.07 to 1.86). CONCLUSIONS Consistent findings across case-control and cohort investigations, complemented by pooled analyses, indicated important roles for anthropometric and hormonal risk factors in the etiology of male breast cancer. Further investigation should focus on potential roles of endogenous hormones.

Effects of two low-dose oral contraceptives on serum lipids and lipoproteins: Differential changes in high-density lipoprotein subclasses

Oral contraceptives containing DL-norgestrel or norethindrone with ethinyl estradiol were administered by random assignment to 21 menstruating women, matched for anthropometric measurements, age, diet, alcohol consumption, smoking, and exercise habits. Pretreatment and 7-week treatment blood samples were obtained and assayed for serum cholesterol, triglyceride high-density lipoprotein cholesterol (HDL-C), and total high-density lipoprotein (HDL), HDL2a, HDL2b, and HDL3 subclasses by analytic ultracentrifugation. Subjects using the norethindrone oral contraceptive had a significant increase in HDL-C: baseline, 46 mg/dl; 7 weeks, 51 mg/dl. Values for the subjects using the norgestrel oral contraceptive were not significantly changed: 46 and 44 mg/dl, respectively. Users of the norethindrone oral contraceptive had significant elevations of total HDL and HDL3, while norgestrel oral contraceptive users demonstrated no significant changes. HDL2b increased with the norethindrone oral contraceptive and declined with the norgestrel oral contraceptive. The changes in HDL2b from baseline to treatment were not significant (p greater than 0.05), but the change with the norethindrone oral contraceptive did differ significantly from that with the norgestrel oral contraceptive (p less than 0.02). These changes may indicate oral contraceptive-induced alterations in HDL structure and metabolism that could be related to the risk of development of atherosclerosis.

Some suggested improvements to current statistical methods of analyzing contraceptive efficacy

4 statistical methods for analyzing the efficacy of contraceptives are presented. The methods explained include: 1) the Tietze-Potter method 2) the Chiang method 3) the exponential method and 4) the log-rank method. Assumptions for each method are presented and notation standardized so that the methods may be uniformly compared. Under the assumption of each method estimation of probabilities were determined. Comparison of 2 IUDs relative to the use-related risks were made. The log-rank method which has the same assumptions as the other methods but is based on 1-day intervals gives the best estimate of the true probabilities. Life-table methods are essentially in agreement with results using the log-rank method. It is recommended that net rather than crude probabilities be used in computing efficacy. Confidence intervals should be avoided. Tests of hypothesis should be performed using the log-rank chi-square test statistic.

Characteristics Relating to Ovarian Cancer Risk: Collaborative Analysis of 12 U.S. Case-Control Studies. VI. Nonepithelial Cancers among Adults

Nonepithelial ovarian cancers are rare, and little is known about their etiology. Of particular interest are the effects of oral contraceptive use and pregnancy, both of which are associated with large decreases in risk for epithelial ovarian cancer. We examined the risk factors for nonepithelial ovarian tumors by combining data from four case-control studies conducted in the United States. We compared personal characteristics of 38 germ cell cases and 45 stromal cases, respectively, with 1,142 and 2,617 general population controls. All subjects were over age 18 years. For germ cell tumors, there was a weak negative association with parity but no consistent pattern of decreasing risk with increasing parity. In contrast, relative to nulligravid women, gravid nulliparous women were at increased risk of developing a germ cell cancer [odds ratio (OR) = 4.8, 95% confidence interval (CI) = 1.2-18.61. The use of oral contraceptives was also associated with elevated risk (OR = 2.0,95% CI = 0.77- 5.1); however, no clear trends in risk were observed. For stromal tumors, oral contraceptive use was associated with decreased risk (OR = 0.37, 95% CI = 0.16-0.83), whereas pregnancy was associated with a small elevation in risk. A trend of increasing risk with increasing age at first term pregnancy was observed, with an odds ratio of 3.6 (95% CI = 1.0-12.5) for a first birth after age 29 years. Risk factors for nonepithelial ovarian cancers do not appear to parallel each other or those for epithelial ovarian cancer. (Epidemiology 1992;3:490-495)