Luis Carvalho

Accurate glycopeptide identification in mass spectrometry-based glycoproteomics is a challenging problem at scale. Recent innovation has been made in increasing the scope and accuracy of glycopeptide identifications, with more precise uncertainty estimates for each part of the structure. We present a dynamically adapting relative retention time model for detecting and correcting ambiguous glycan assignments that are difficult to detect from fragmentation alone, a layered approach to glycopeptide fragmentation modeling that improves N-glycopeptide identification in samples without compromising identification quality, and a site-specific method to increase the depth of the glycoproteome confidently identifiable even further. We demonstrate our techniques on a set of previously published datasets, showing the performance gains at each stage of optimization. These techniques are provided in the open-source glycomics and glycoproteomics platform GlycReSoft available at https://github.com/mobiusklein/glycresoft.

Model evaluation is of crucial importance in modern statistics application. The construction of ROC and calculation of AUC have been widely used for binary classification evaluation. Recent research generalizing the ROC/AUC analysis to multi-class classification has problems in at least one of the four areas: 1. failure to provide sensible plots 2. being sensitive to imbalanced data 3. unable to specify mis-classification cost and 4. unable to provide evaluation uncertainty quantification. Borrowing from a binomial matrix factorization model, we provide an evaluation metric summarizing the pair-wise multi-class True Positive Rate (TPR) and False Positive Rate (FPR) with one-dimensional vector representation. Visualization on the representation vector measures the relative speed of increment between TPR and FPR across all the classes pairs, which in turns provides a ROC plot for the multi-class counterpart. An integration over those factorized vector provides a binary AUC-equivalent summary on the classifier performance. Mis-clasification weights specification and bootstrapped confidence interval are also enabled to accommodate a variety of of evaluation criteria. To support our findings, we conducted extensive simulation studies and compared our method to the pair-wise averaged AUC statistics on benchmark datasets.

We investigate a general matrix factorization for deviance-based data losses, extending the ubiquitous singular value decomposition beyond squared error loss. While similar approaches have been explored before, our method leverages classical statistical methodology from generalized linear models (GLMs) and provides an efficient algorithm that is flexible enough to allow for structural zeros via entry weights. Moreover, by adapting results from GLM theory, we provide support for these decompositions by (i) showing strong consistency under the GLM setup, (ii) checking the adequacy of a chosen exponential family via a generalized Hosmer-Lemeshow test, and (iii) determining the rank of the decomposition via a maximum eigenvalue gap method. To further support our findings, we conduct simulation studies to assess robustness to decomposition assumptions and extensive case studies using benchmark datasets from image face recognition, natural language processing, network analysis, and biomedical studies. Our theoretical and empirical results indicate that the proposed decomposition is more flexible, general, and robust, and can thus provide improved performance when compared to similar methods. To facilitate applications, an R package with efficient model fitting and family and rank determination is also provided.

The Hansen et al . critique centers on the lack of spatial agreement between two very different datasets. Nonetheless, properly constructed comparisons designed to reconcile the two datasets yield up to 90% agreement (e.g., in South America).

Estimation of origin–destination (OD) demand plays a key role in successful transportation studies. In this paper, we consider the estimation of time-varying day-to-day OD flows given data on traffic volumes in a transportation network for a sequence of days. We propose a dynamic linear model (DLM) in order to represent the stochastic evolution of OD flows over time. DLMs are Bayesian state-space models which can capture non-stationarity. We take into account the hierarchical relationships between the distribution of OD flows among routes and the assignment of traffic volumes on links. Route choice probabilities are obtained through a utility model based on past route costs. We propose a Markov chain Monte Carlo algorithm, which integrates Gibbs sampling and a forward filtering backward sampling technique, in order to approximate the joint posterior distribution of mean OD flows and parameters of the route choice model. Our approach can be applied to congested networks and in the case when data are available on only a subset of links. We illustrate the application of our approach through simulated experiments on a test network from the literature.

Forests out of balance: Are tropical forests a net source or net sink of atmospheric carbon dioxide? As fundamental a question as that is, there still is no agreement about the answer, with different studies suggesting that it is anything from a sizable sink to a modest source. Baccini et al. used 12 years of MODIS satellite data to determine how the aboveground carbon density of woody, live vegetation has changed throughout the entire tropics on an annual basis. They find that the tropics are a net carbon source, with losses owing to deforestation and reductions in carbon density within standing forests being double that of gains resulting from forest growth. Science , this issue p. 230

As both a central task in Remote Sensing and a common problem in many other situations involving time series data, change point detection boasts a thorough and well-documented history of study. However, the treatment of missing data and proper exploitation of the structure in multivariate time series during change point detection remains lacking. Multispectral, high temporal resolution time series data from NASA's Moderate Resolution Imaging Spectroradiometer (MODIS) instruments provide an attractive and challenging context to contribute to the change point detection literature. In an effort to better monitor change in land cover using MODIS data, we present a novel approach to identifying periods of time in which regions experience some conversion-type of land cover change. That is, we propose a method for parameter estimation and change point detection in the presence of missing data which capitalizes on the high dimensionality of MODIS data. We test the quality of our method in a simulation study alongside a contemporary change point method and apply it in a case study at the Xingu River Basin in the Amazon. Not only does our method maintain a high accuracy, but can provide insight into the types of changes occurring via land cover conversion probabilities. In this way we can better characterize the amount and types of forest disturbance in our study area in comparison to traditional change point methods.