Edward J. Bedrick

Comparison of multi-subject ICA methods for analysis of fMRI data.

Spatial independent component analysis (ICA) applied to functional magnetic resonance imaging (fMRI) data identifies functionally connected networks by estimating spatially independent patterns from their linearly mixed fMRI signals. Several multi‐subject ICA approaches estimating subject‐specific time courses (TCs) and spatial maps (SMs) have been developed, however, there has not yet been a full comparison of the implications of their use. Here, we provide extensive comparisons of four multi‐subject ICA approaches in combination with data reduction methods for simulated and fMRI task data. For multi‐subject ICA, the data first undergo reduction at the subject and group levels using principal component analysis (PCA). Comparisons of subject‐specific, spatial concatenation, and group data mean subject‐level reduction strategies using PCA and probabilistic PCA (PPCA) show that computationally intensive PPCA is equivalent to PCA, and that subject‐specific and group data mean subject‐level PCA are preferred because of well‐estimated TCs and SMs. Second, aggregate independent components are estimated using either noise‐free ICA or probabilistic ICA (PICA). Third, subject‐specific SMs and TCs are estimated using back‐reconstruction. We compare several direct group ICA (GICA) back‐reconstruction approaches (GICA1‐GICA3) and an indirect back‐reconstruction approach, spatio‐temporal regression (STR, or dual regression). Results show the earlier group ICA (GICA1) approximates STR, however STR has contradictory assumptions and may show mixed‐component artifacts in estimated SMs. Our evidence‐based recommendation is to use GICA3, introduced here, with subject‐specific PCA and noise‐free ICA, providing the most robust and accurate estimated SMs and TCs in addition to offering an intuitive interpretation. Hum Brain Mapp, 2011.

Identification of novel cluster groups in pediatric high-risk B-precursor acute lymphoblastic leukemia with gene expression profiling: correlation with genome-wide DNA copy number alterations, clinical characteristics, and outcome

To resolve the genetic heterogeneity within pediatric high-risk B-precursor acute lymphoblastic leukemia (ALL), a clinically defined poor-risk group with few known recurring cytogenetic abnormalities, we performed gene expression profiling in a cohort of 207 uniformly treated children with high-risk ALL. Expression profiles were correlated with genome-wide DNA copy number abnormalities and clinical and outcome features. Unsupervised clustering of gene expression profiling data revealed 8 unique cluster groups within these high-risk ALL patients, 2 of which were associated with known chromosomal translocations (t(1;19)(TCF3-PBX1) or MLL), and 6 of which lacked any previously known cytogenetic lesion. One unique cluster was characterized by high expression of distinct outlier genes AGAP1, CCNJ, CHST2/7, CLEC12A/B, and PTPRM; ERG DNA deletions; and 4-year relapse-free survival of 94.7% ± 5.1%, compared with 63.5% ± 3.7% for the cohort (P = .01). A second cluster, characterized by high expression of BMPR1B, CRLF2, GPR110, and MUC4; frequent deletion of EBF1, IKZF1, RAG1-2, and IL3RA-CSF2RA; JAK mutations and CRLF2 rearrangements (P < .0001); and Hispanic ethnicity (P < .001) had a very poor 4-year relapse-free survival (21.0% ± 9.5%; P < .001). These studies reveal striking clinical and genetic heterogeneity in high-risk ALL and point to novel genes that may serve as new targets for diagnosis, risk classification, and therapy.

A New Perspective on Priors for Generalized Linear Models

Abstract This article deals with specifications of informative prior distributions for generalized linear models. Our emphasis is on specifying distributions for selected points on the regression surface; the prior distribution on regression coefficients is induced from this specification. We believe that it is inherently easier to think about conditional means of observables given the regression variables than it is to think about model-dependent regression coefficients. Previous use of conditional means priors seems to be restricted to logistic regression with one predictor variable and to normal theory regression. We expand on the idea of conditional means priors and extend these to arbitrary generalized linear models. We also consider data augmentation priors where the prior is of the same form as the likelihood. We show that data augmentation priors are special cases of conditional means priors. With current Monte Carlo methodology, such as importance sampling and Gibbs sampling, our priors result in...

Model Selection for Multivariate Regression in Small Samples

We develop a small-sample criterion (AICc) for selecting multivariate regression models. This criterion adjusts the Akaike information criterion to be an exact unbiased estimator for the expected Kullback-Leibler information. A small-sample comparison shows that AICC provides better model order choices than other available model selection methods. Data from an agricultural experiment are analyzed.

GENE EXPRESSION CLASSIFIERS FOR RELAPSE FREE SURVIVAL AND MINIMAL RESIDUAL DISEASE IMPROVE RISK CLASSIFICATION AND OUTCOME PREDICTION IN PEDIATRIC B-PRECURSOR ACUTE LYMPHOBLASTIC LEUKEMIA

To determine whether gene expression profiling could improve outcome prediction in children with acute lymphoblastic leukemia (ALL) at high risk for relapse, we profiled pretreatment leukemic cells in 207 uniformly treated children with high-risk B-precursor ALL. A 38-gene expression classifier predictive of relapse-free survival (RFS) could distinguish 2 groups with differing relapse risks: low (4-year RFS, 81%, n = 109) versus high (4-year RFS, 50%, n = 98; P < .001). In multivariate analysis, the gene expression classifier (P = .001) and flow cytometric measures of minimal residual disease (MRD; P = .001) each provided independent prognostic information. Together, they could be used to classify children with high-risk ALL into low- (87% RFS), intermediate- (62% RFS), or high- (29% RFS) risk groups (P < .001). A 21-gene expression classifier predictive of end-induction MRD effectively substituted for flow MRD, yielding a combined classifier that could distinguish these 3 risk groups at diagnosis (P < .001). These classifiers were further validated on an independent high-risk ALL cohort (P = .006) and retainedindependent prognostic significance (P < .001) in the presence of other recently described poor prognostic factors (IKAROS/IKZF1 deletions, JAK mutations, and kinase expression signatures). Thus, gene expression classifiers improve ALL risk classification and allow prospective identification of children who respond or fail current treatment regimens. These trials were registered at http://clinicaltrials.gov under NCT00005603.

Lung toxicity of ambient particulate matter from southeastern U.S. sites with different contributing sources : Relationships between composition and effects

Background Exposure to air pollution and, more specifically, particulate matter (PM) is associated with adverse health effects. However, the specific PM characteristics responsible for biological effects have not been defined. Objectives In this project we examined the composition, sources, and relative toxicity of samples of PM with aerodynamic diameter ≥2.5 μm (PM2.5) collected from sites within the Southeastern Aerosol Research and Characterization (SEARCH) air monitoring network during two seasons. These sites represent four areas with differing sources of PM2.5, including local urban versus regional sources, urban areas with different contributions of transportation and industrial sources, and a site influenced by Gulf of Mexico weather patterns. Methods We collected samples from each site during the winter and summer of 2004 for toxicity testing and for chemical analysis and chemical mass balance–based source apportionment. We also collected PM2.5 downwind of a series of prescribed forest burns. We assessed the toxicity of the samples by instillation into rat lungs and assessed general toxicity, acute cytotoxicity, and inflammation. Statistical dose–response modeling techniques were used to rank the relative toxicity and compare the seasonal differences at each site. Projection-to-latent-surfaces (PLS) techniques examined the relationships among sources, chemical composition, and toxicologic end points. Results and conclusions Urban sites with high contributions from vehicles and industry were most toxic.

Changing Relationship of Blood Pressure with Mortality over Time among Hemodialysis Patients

High BP is a major risk factor for atherosclerotic cardiovascular disease mortality in the general population. Surprising, studies that have been conducted among hemodialysis (HD) patients have yielded conflicting data on the relationship between BP and mortality. This study explores two hypotheses among HD patients: (1) The relationship between BP and mortality changes over time, and (2) mild to moderate hypertension is well tolerated. Incident HD patients who were treated at Dialysis Clinic Inc. facilities between 1993 and 2003 were studied. Primary end points were atherosclerotic cardiovascular disease and all-cause mortality. The relationship between BP and mortality was analyzed in two sets of Cox proportional hazards models. Model-B explored the relationship between baseline BP and mortality in sequential time periods. Model-TV assessed the relationship between BP, treated as time-varying, and mortality. The study sample (n = 16,959) was similar in characteristics to the United States Renal Data Systems population, although black patients were slightly overrepresented. Model-B demonstrated that the relationship between baseline BP and mortality changes over time. Low systolic BP (<120 mmHg) was associated with increased mortality in years 1 and 2. High systolic BP (> or =150 mmHg) was associated with increased mortality among patients who survived > or =3 yr. Low pulse pressure was associated with increased mortality. Model-TV demonstrated that mild to moderate systolic hypertension may be relatively well tolerated. In conclusion, the relationship between baseline BP and mortality changes over time. Mild to moderate systolic hypertension was associated with only modest increases in mortality.

Which Targets in Clinical Practice Guidelines Are Associated with Improved Survival in a Large Dialysis Organization

Professional organizations have developed practice guidelines in the hope of improving clinical outcomes. The National Kidney Foundations Kidney Disease Outcomes Quality Initiative (KDOQI) has set targets for dialysis dosage (single-pool Kt/V), hematocrit, serum albumin, calcium, phosphorus, parathyroid hormone, and BP for hemodialysis (HD) patients. Several guidelines are largely based on results from observational studies. In contrast to other parameters, BP values within the KDOQI guidelines have been associated with increased mortality. Therefore, it was postulated that having multiple parameters that satisfy the current guidelines, except those for BP, is associated with improved survival among HD patients. A retrospective analysis was conducted of incident HD patients who were treated at facilities operated by Dialysis Clinic Inc., a not-for-profit dialysis provider, between January 1, 1998, and December 31, 2004 (n = 13,792). Cox proportional hazards models were used to assess the association between satisfying guidelines and mortality. Values within guidelines for single-pool Kt/V, hematocrit, serum albumin, calcium, phosphorus, and parathyroid hormone were associated with decreased mortality (P < or = 0.0001). The largest survival benefit was found for serum albumin (hazard ratio [HR] 0.27; 95% confidence interval [CI] 0.24 to 0.31). Satisfying these six guidelines simultaneously was associated with an 89% reduction in mortality (HR 0.11; 95% CI 0.06 to 0.19]). Conversely, BP values satisfying the guideline were associated with increased mortality (HR 1.90; 95% CI 1.73 to 2.10). Because this target was largely extrapolated from the general population, a randomized, controlled trial is needed to identify the optimal BP for HD patients.

Sequential actions of myotubularin lipid phosphatases regulate endosomal PI(3)P and growth factor receptor trafficking.

Two different human diseases, X-linked myotubular myopathy and Charcot-Marie-Tooth disease, result from mutant MTM1 or MTMR2 lipid phosphatases. Although events involved in endosomal PI(3)P and PI(3,5)P(2) synthesis are well established and pivotal in receptor signaling and degradation, enzymes involved in phosphoinositide degradation and their roles in trafficking are incompletely characterized. Here, we dissect the functions of the MTM1 and MTMR2 myotubularins and establish how they contribute to endosomal PI(3)P homeostasis. By mimicking loss of function in disease through siRNA-mediated depletion of the myotubularins, excess PI(3)P accumulates on early (MTM1) and late (MTMR2) endosomes. Surprisingly, the increased PI(3)P blocks the egress of epidermal growth factor receptors from early or late endosomes, suggesting that the accumulation of signaling receptors in distinct endosomes may contribute to the unique disease etiologies when MTM1 or MTMR2 are mutant. We further demonstrate that direct myotubularin binding to the type III PI 3-kinase complex hVps34/hVps15 leads to phosphatase inactivation. The lipid kinase-phosphatase interaction also precludes interaction of the PI 3-kinase with Rab GTPase activators. Thus, unique molecular complexes control kinase and phosphatase activation and locally regulate PI(3)P on discrete endosome populations, thereby providing a molecular rationale for related human myo- and neuropathies.

A conserved signal and GTPase complex are required for the ciliary transport of polycystin-1.

Ciliary delivery of polycystin-1 depends on a conserved (K/R/Q)VxPx motif. The signal enables Arf4 GTPase binding and assembly of a multimeric trafficking complex. Functional importance of Arf4 and Rab8 in ciliary trafficking is shown. The studies offer the first unifying molecular rationale for human cystic kidney diseases and retinopathies.