More Powerful Selective Inference for the Graph Fused Lasso

The graph fused lasso-which includes as a special case the one-dimensional fused lasso-is widely used to reconstruct signals that are piecewise constant on a graph, meaning that nodes connected by an edge tend to have identical values. We consider testing for a difference in the means of two connected components estimated using the graph fused lasso. A naive procedure such as a z-test for a difference in means will not control the selective Type I error, since the hypothesis that we are testing is itself a function of the data. In this work, we propose a new test for this task that controls the selective Type I error, and conditions on less information than existing approaches, leading to substantially higher power. We illustrate our approach in simulation and on datasets of drug overdose death rates and teenage birth rates in the contiguous United States. Our approach yields more discoveries on both datasets. Supplementary materials for this article are available online.

Testing for a Change in Mean After Changepoint Detection

While many methods are available to detect structural changes in a time series, few procedures are available to quantify the uncertainty of these estimates post-detection. In this work, we fill this gap by proposing a new framework to test the null hypothesis that there is no change in mean around an estimated changepoint. We further show that it is possible to efficiently carry out this framework in the case of changepoints estimated by binary segmentation and its variants, ℓ 0 segmentation, or the fused lasso. Our setup allows us to condition on much less information than existing approaches, which yields higher powered tests. We apply our proposals in a simulation study and on a dataset of chromosomal guanine-cytosine content. These approaches are freely available in the R package ChangepointInference at https://jewellsean.github.io/changepoint-inference/.

It's complicated: characterizing the time-varying relationship between cell phone mobility and COVID-19 spread in the US

Restricting in-person interactions is an important technique for limiting the spread of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). Although early research found strong associations between cell phone mobility and infection spread during the initial outbreaks in the United States, it is unclear whether this relationship persists across locations and time. We propose an interpretable statistical model to identify spatiotemporal variation in the association between mobility and infection rates. Using 1 year of US county-level data, we found that sharp drops in mobility often coincided with declining infection rates in the most populous counties in spring 2020. However, the association varied considerably in other locations and across time. Our findings are sensitive to model flexibility, as more restrictive models average over local effects and mask much of the spatiotemporal variation. We conclude that mobility does not appear to be a reliable leading indicator of infection rates, which may have important policy implications.

Inferring spikes from calcium imaging in dopamine neurons

Calcium imaging has led to discoveries about neural correlates of behavior in subcortical neurons, including dopamine (DA) neurons. However, spike inference methods have not been tested in most populations of subcortical neurons. To address this gap, we simultaneously performed calcium imaging and electrophysiology in DA neurons in brain slices and applied a recently developed spike inference algorithm to the GCaMP fluorescence. This revealed that individual spikes can be inferred accurately in this population. Next, we inferred spikes in vivo from calcium imaging from these neurons during Pavlovian conditioning, as well as during navigation in virtual reality. In both cases, we quantitatively recapitulated previous in vivo electrophysiological observations. Our work provides a validated approach to infer spikes from calcium imaging in DA neurons and implies that aspects of both tonic and phasic spike patterns can be recovered.

A large-scale standardized physiological survey reveals functional organization of the mouse visual cortex

To understand how the brain processes sensory information to guide behavior, we must know how stimulus representations are transformed throughout the visual cortex. Here we report an open, large-scale physiological survey of activity in the awake mouse visual cortex: the Allen Brain Observatory Visual Coding dataset. This publicly available dataset includes the cortical activity of nearly 60,000 neurons from six visual areas, four layers, and 12 transgenic mouse lines in a total of 243 adult mice, in response to a systematic set of visual stimuli. We classify neurons on the basis of joint reliabilities to multiple stimuli and validate this functional classification with models of visual responses. While most classes are characterized by responses to specific subsets of the stimuli, the largest class is not reliably responsive to any of the stimuli and becomes progressively larger in higher visual areas. These classes reveal a functional organization wherein putative dorsal areas show specialization for visual motion signals.

Somatic mutation detection and classification through probabilistic integration of clonal population information

Somatic mutations are a primary contributor to malignancy in human cells. Accurate detection of mutations is needed to define the clonal composition of tumours whereby clones may have distinct phenotypic properties. Although analysis of mutations over multiple tumour samples from the same patient has the potential to enhance identification of clones, few analytic methods exploit the correlation structure across samples. We posited that incorporating clonal information into joint analysis over multiple samples would improve mutation detection, particularly those with low prevalence. In this paper, we develop a new procedure called MuClone, for detection of mutations across multiple tumour samples of a patient from whole genome or exome sequencing data. In addition to mutation detection, MuClone classifies mutations into biologically meaningful groups and allows us to study clonal dynamics. We show that, on lung and ovarian cancer datasets, MuClone improves somatic mutation detection sensitivity over competing approaches without compromising specificity.

EXACT SPIKE TRAIN INFERENCE VIA ℓ0 OPTIMIZATION

In recent years new technologies in neuroscience have made it possible to measure the activities of large numbers of neurons simultaneously in behaving animals. For each neuron a fluorescence trace is measured; this can be seen as a first-order approximation of the neuron's activity over time. Determining the exact time at which a neuron spikes on the basis of its fluorescence trace is an important open problem in the field of computational neuroscience. Recently, a convex optimization problem involving an ℓ1 penalty was proposed for this task. In this paper we slightly modify that recent proposal by replacing the ℓ1 penalty with an ℓ0 penalty. In stark contrast to the conventional wisdom that ℓ0 optimization problems are computationally intractable, we show that the resulting optimization problem can be efficiently solved for the global optimum using an extremely simple and efficient dynamic programming algorithm. Our R-language implementation of the proposed algorithm runs in a few minutes on fluorescence traces of 100,000 timesteps. Furthermore, our proposal leads to substantial improvements over the previous ℓ1 proposal, in simulations as well as on two calcium imaging datasets. R-language software for our proposal is available on CRAN in the package LZeroSpikeInference. Instructions for running this software in python can be found at https://github.com/jewellsean/LZeroSpikeInference.

Microsatellite instability predicts improved response to adjuvant therapy with irinotecan, 5-fluorouracil and leucovorin in stage III colon cancer

10003

Background: Colon cancers exhibiting a high level of microsatellite instability (MSI-H) show distinct clinicopathological features, including both better prognosis and reduced response to 5-fluorouracil (5-FU)-based chemotherapy. We investigated the impact of adjuvant chemotherapy containing irinotecan in patients with MSI-H colon cancers. Methods: CALGB protocol 89803 randomized 1264 patients with resected stage III colon cancer to receive post-operative 5-FU and leucovorin (LV) with or without irinotecan. Paraffin blocks containing primary tumor and normal tissue were collected. Microsatellite instablility was assessed using a panel of mono- and di-nucleotide markers. Disease free survival (DFS) was measured from trial entry until documented disease progression or death from any cause. A statistical significance level of 0.2 was used in screening to generate hypotheses regarding MSI status and outcome. Median follow-up at analysis was 3.8 years. Overall C89803 showed no advantage for addition of irinotecan to 5-FU/LV. Results: Patients with and without tumor samples analyzed did not differ by treatment, age, gender, primary site, T-stage, differentiation, # positive nodes, or mucinous type. Of 482 tumors analyzed, 75 (16%) demonstrated MSI-H. MSI-H cancers were more likely to be located in the proximal colon (p<0.0001), of high histologic grade (p<0.0001) and mucinous histology (p<0.0001), and also had increased numbers of tumor-containing lymph nodes (mean # positive nodes/case = 3.5 for MSI Low/Stable vs. 4.7 for MSI-H; p = 0.04). At the time of analysis 143 of 482 patients (36%) analyzed experienced tumor recurrence and/or death due to any cause. For patients with MSI-H tumors, DFS was better in those treated with irinotecan in addition to 5-FU/LV (logrank p=0.18). Among patients with MSI Low/Stable tumors there was no difference in DFS between those treated with and without irinotecan (logrank p =0.39). Conclusions: Early results from CALGB protocol 89803 indicate that addition of postoperative irinotecan to 5-FU/LV may improve DFS in patients with stage III colon cancers that exhibit MSI-H. Longer follow-up is required to confirm this finding.

[Table: see text]

Optical characterization of a dual-frequency hybrid aligned nematic liquid crystal cell

The dielectric anisotropy of a highly dispersive dual-frequency nematic liquid crystal (MDA-00-3969 (Merck KGa)) has been determined using the optical fully-leaky guided-mode technique. A 4Vrms sinusoidal voltage was applied across a 5microm hybrid aligned nematic (HAN) cell at various frequencies in both the positive and negative dielectric anisotropy regime. Optical data was collected at each frequency enabling the director profile in each case to be determined using a multi-layer optics model in combination with a liquid crystal free-energy minimization routine. The thresholdless response of the HAN cell combined with the extreme sensitivity of the optical characterization technique has allowed subtle changes in dielectric permittivity with frequency to be observed. The resulting measured dispersion shows excellent agreement with a single Debye-type relaxation model.

A randomized trial of 72- versus 24-hour intravenous tubing set changes in newborns receiving lipid therapy

OBJECTIVE

To compare the microbial contamination rate of infusate in the intravenous tubing of newborns receiving lipid therapy, replacing the intravenous delivery system at 72-hour versus 24-hour intervals.

DESIGN

Infants requiring intravenous lipid therapy were randomly assigned to have intravenous sets changed on a 72- or a 24-hour schedule, in a 3:1 ratio, in order to compare the infusate contamination rates in an equivalent number of tubing sets.

SETTING

A 35-bed, teaching, referral, neonatal intensive-care unit (NICU).

PARTICIPANTS

All neonates admitted to the NICU for whom intravenous lipid was ordered.

METHODS

Patients were randomized in pharmacy, on receipt of the order for intravenous lipid therapy, to either 72- or 24-hour administration set changes, and followed until 1 week after discontinuation of lipids or discharge from the NICU. Microbial contamination of the infusate was assessed in both groups at the time of administration set changes. Contamination rates were analyzed separately for the lipid and amino acid-glucose tubing sets. Patient charts were reviewed for clinical and epidemiological data, including birth weight, gestational age, gender, age at start of lipid therapy, duration of parenteral nutrition, and type of intravenous access.

RESULTS

During the study period, 1,101 and 1,112 sets were sampled in the 72- and 24-hour groups, respectively. Microbial contamination rates were higher in the 72-hour group than the 24-hour group for lipid infusions (39/1,101 [3.54%] vs 15/1,112 [1.35%]; P=.001) and for amino acid infusions (12/1,093 [1.10%] vs 4/1,103 [0.36%]; P=.076). Logistic regression analysis controlling for birth weight, gestational age, and type of venous access showed that only the tubing change interval was significantly associated with lipid set contaminations (odds ratio, 2.69; P=.0013). The rate of blood cultures ordered was higher in the 72- versus the 24-hour group (6.11 vs 4.99 per 100 patient days of total parenteral nutrition; P=.017), and a higher proportion of infants randomized to the 72-hour group died (8% vs 4%; P=.05), although the excess deaths could not clearly be attributed to bacteremia.

CONCLUSION

Microbial contamination of infusion sets is significantly more frequent with 72- than with 24-hour set changes in neonates receiving lipid solutions. This may be associated with an increased mortality rate.

Elderly patients' participation in discharge decision making: 2

This article is the second in a two-part series describing a study that explored patient participation in discharge decision making. The study focused on the participatory relationship between older patients and nurses and was performed in two phases. As reported in the previous article (Vol 5(15): 914-16, 929-32) the first phase involved group interviews with primary nurses from two elderly care units. This article reports the findings of the second phase which involved a series of interviews with five elderly patients during their hospital admission. The data from both phases of the study were analysed using the software package Ethnograph. Unlike nurses, the patients in this study had little understanding of the term 'participation'. They were mainly concerned with being given information and reported having unmet information needs. They felt that they did have a role in discharge decision making but were unable to fulfil it, consequently they relied upon nursing staff to take the lead in the decision-making process.

Elderly patients' participation in discharge decision making: 1

This article is the first of a two part series describing a study that explored patient participation in discharge decision making. The study involved two phases and focused on the participatory relationship between older patients and nurses. In phase one, the author conducted a group interview with primary nurses at two elderly care units specializing in rehabilitation. The second phase involved a series of interviews with five elderly patients during their hospital admission. The data from both phases of the study were analysed using the Ethnograph software package. This first article discusses the views of the nurses while the views of patients are addressed in the second article. Phase one of the study showed that the nurses interviewed had an elaborate understanding of participation. They tended to stress practical examples of participation rather than stating the reasons why it should be encouraged. Informants recognized the need to involve patients in decisions about their discharge from hospital. However, they viewed elderly patients as being both passive and reluctant to participate in decision making. Several reasons were given for this which were seen as resulting from patient rather than nursing behaviour. Part two of this article will appear in Vol 5(17): 26 September.

A brief, telephone-administered food frequency questionnaire can be useful for surveillance of dietary fat intakes

A 13-item questionnaire designed for quick telephone administration was evaluated for use in surveillance of fat intake in the United States. Study populations included 560 middle-aged and older adults from Beaver Dam, WI, 252 middle-aged and older women from Wisconsin, 73 young, low income Hispanic women from Chicago, IL, 52 older adults from Arizona and 135 younger adults from Augusta, GA. Correlations between fat scores and fat intakes measured by multiple food records or recalls or by more extensive food frequency questionnaires ranged from 0.33 to 0.60, similar to results from other published questionnaire validation studies. Correlations with percentage of energy from fat were lower (0.26 to 0.42), except for the Chicago population, for which there was no correlation (-0.02). There was no systematic variation in correlations among other subgroups defined by demographic and health-related characteristics, including race (black vs. white). Most, but not all, of the substantial differences in fat intakes among subgroups were identified by the questionnaire. The questionnaire will not capture small differences in intakes among groups and is inappropriate when the sample size is limited or for populations with diets substantially different from the typical U.S. diets, such as the Chicago population. However, with attention to its limitations, the questionnaire is useful for surveillance.

Evaluation of a brief telephone questionnaire to estimate fruit and vegetable consumption in diverse study populations

We evaluated the use of a six-item telephone questionnaire to estimate fruit and vegetable intakes in five diverse populations. Researchers administered the telephone questionnaire to persons who had previously undergone more extensive dietary assessment. The study population included 553 middle-aged and older adults in Beaver Dam, WI; 252 middle-aged and older women throughout Wisconsin; 150 parents of school children in Augusta, GA; 73 low-income, Hispanic mothers in Chicago, IL; and 51 older adults in Arizona. Spearman correlation coefficients between total fruit and vegetable intakes measured by the brief telephone survey and by more extensive food frequency questionnaires were 0.47 (Augusta), 0.48 (Arizona), 0.56 (Wisconsin), and 0.57 (Beaver Dam). Correlations between intakes measured by the brief telephone survey and by multiple diet records or recalls were 0.29 (Arizona), 0.46 (Chicago), and 0.54 (Beaver Dam). With the exception of Arizona, mean daily fruit and vegetable intakes measured by the telephone survey were similar to intakes estimated by multiple diet records or recalls and lower than those estimated by extensive food frequency questionnaires. Although caution may be needed in interpreting dietary reports from some ethnic subgroups, this brief telephone questionnaire may be useful for surveillance of fruit and vegetable intake in the United States.

The accuracy of parental reports of their children's intake of fruits and vegetables: validation of a food frequency questionnaire with serum levels of carotenoids and vitamins C, A, and E

It has been recommended that U.S. children increase their dietary intake of fruits and vegetables. Measuring diets of children to support and evaluate nutritional interventions can be a difficult task, however. We administered to 97 parents of children age 6-10 years a food frequency questionnaire on their children's usual dietary intake over the previous 3 months. We then compared these reports by parents of their children's intakes of fruits and vegetables, and the derived estimates of intake of carotenoids and vitamins C, A, and E, with the children's serum levels of carotenoids and vitamins C, A, and E. The dietary reports of intakes of 35 fruits and vegetables showed Spearman rank-order correlations of 0.30 with serum carotenoids and 0.34 with serum vitamin C. Children in the highest quartile for intake of fruits and vegetables according to their parents' food frequency reports had 35% higher carotene levels and 31% higher vitamin C levels in their serum than did children in the lowest quartile for intake of fruits and vegetables. We conclude that parental reports of young children's diets using food frequency methods are accurate enough to be useful in nutritional screening and dietary surveillance of fruit and vegetable intake.

Suppression of experimental autoimmune encephalomyelitis by gallium nitrate

We examined the effect of gallium (Ga) nitrate on the development of the development of experimental autoimmune encephalomyelitis (EAE). Weekly subcutaneous injections of 10-30 mg/kg prevented clinical signs as well as histopathological changes of EAE. The optimal timing of a single injection of Ga was 6 days after induction of EAE, with amelioration also apparent following a single injection on day 3 or 9 but not day 12. Ga administered in vivo suppressed myelin basic protein (MBP) and purified protein derivative-specific lymphocyte proliferative responses in vitro. Addition of Ga to MBP-specific T lymphocyte line cultures at various times after initiation of culture revealed that Ga exerts an effect at an early stage of cellular activation.

Thiazide therapy and severe hypercalcemia in a patient with hyperparathyroidism

Images

Liquid-chromatographic determination of total hydroxyproline in urine

In this method for quantifying 4-hydroxyproline in human urine, 50 microL of urine is hydrolyzed, derivatized with phenylisothiocyanate (PITC), and then quantified by reversed-phase HPLC with ultraviolet detection. The detection limit in urine is 373 pg of hydroxyproline per 50-microL injection. The total CVs for high- and low-concentration pools are 5.3% and 3.9%, respectively (10 runs in 10 days). The standard curve of the assay is linear over a range of 0 to 22 nmol per injection. We estimated the normal range for hydroxyproline excretion in men on an unrestricted diet to be 123-308 mumol/24 h. We also report hydroxyproline concentrations in patients with metastatic bone disease and cirrhosis of the liver.

Liquid-chromatographic determination of total hydroxyproline in urine

In this method for quantifying 4-hydroxyproline in human urine, 50 microL of urine is hydrolyzed, derivatized with phenylisothiocyanate (PITC), and then quantified by reversed-phase HPLC with ultraviolet detection. The detection limit in urine is 373 pg of hydroxyproline per 50-microL injection. The total CVs for high- and low-concentration pools are 5.3% and 3.9%, respectively (10 runs in 10 days). The standard curve of the assay is linear over a range of 0 to 22 nmol per injection. We estimated the normal range for hydroxyproline excretion in men on an unrestricted diet to be 123-308 mumol/24 h. We also report hydroxyproline concentrations in patients with metastatic bone disease and cirrhosis of the liver.

Limits of sensitivity for the radioimmunodetection of colon cancer by means of a hand held gamma probe

This study was undertaken to define the limits for the radioimmunodetection of minimal deposits of colorectal cancer cells using a hand held gamma probe. 125I labeled monoclonal antibody 17-1A and its F(ab')2 fragments were reacted in vitro with cells of the human colorectal cancer line SW 1116. The limits of sensitivity of the probe were determined by injecting doubling dilutions of 125I-antibody coated SW 1116 cells ranging from 10(7) to 3.9 x 10(4) subserosally at 2 cm intervals into 60 cm segments of freshly obtained autopsy or surgical specimens of human colon. A linear relationship was observed between the number of cells injected and the number of counts obtained with either the probe or well counter. As few as 6.25 x 10(5) 125I-antibody coated cells (less than 1 mm3) were detected under experimentally defined conditions by an earlier version of the probe, and 3.9 x 10(4) coated cells (much less than 1 mm3) could be detected by the currently available model. Although the count rates were less than 5% of those obtained by well counter, nevertheless, these were 10-25 times greater than background and allowed the detection of tumor cell deposits that otherwise would not have been discernible by either palpation or external scintigraphy. These findings, in conjunction with ongoing clinical studies, suggest that the hand held gamma probe may increase the usefulness of monoclonal antibodies for the radioimmunodetection of cancer.