arXiv subjects

William Stafford Noble

Publications and source records attributed to William Stafford Noble.

12 recordsLinked to original sources

arXiv2026

A semi-supervised framework for diverse multiple hypothesis testing scenarios

Standard multiple testing procedures are designed to report a list of discoveries, or suspected false null hypotheses, given the hypotheses' p-values or test scores. Recently there has been a growing interest in enhancing such procedures by combining additional information with the primary p-value or score. In line with this idea, we develop RESET (REScoring via Estimating and Training), which splits the data into a training part and an estimating part so that any semi-supervised learning approach can factor in the available side information while maintaining finite-sample error-rate control. Our practical implementation, RESET Ensemble, selects from an ensemble of classification algorithms so that it is compatible with a range of multiple testing scenarios without the need for the user to select the appropriate one. We apply RESET to both p-value and competition based multiple testing problems and show that RESET is (1) power-wise competitive, (2) fast compared to most tools and (3) able to uniquely achieve finite sample false discovery rate or false discovery exceedance control, depending on the user's preference.

stat.ME↗

arXiv2025

Foundation model for mass spectrometry proteomics

Mass spectrometry is the dominant technology in the field of proteomics, enabling high-throughput analysis of the protein content of complex biological samples. Due to the complexity of the instrumentation and resulting data, sophisticated computational methods are required for the processing and interpretation of acquired mass spectra. Machine learning has shown great promise to improve the analysis of mass spectrometry data, with numerous purpose-built methods for improving specific steps in the data acquisition and analysis pipeline reaching widespread adoption. Here, we propose unifying various spectrum prediction tasks under a single foundation model for mass spectra. To this end, we pre-train a spectrum encoder using de novo sequencing as a pre-training task. We then show that using these pre-trained spectrum representations improves our performance on the four downstream tasks of spectrum quality prediction, chimericity prediction, phosphorylation prediction, and glycosylation status prediction. Finally, we perform multi-task fine-tuning and find that this approach improves the performance on each task individually. Overall, our work demonstrates that a foundation model for tandem mass spectrometry proteomics trained on de novo sequencing learns generalizable representations of spectra, improves performance on downstream tasks where training data is limited, and can ultimately enhance data acquisition and analysis in proteomics experiments.

cs.LG↗

arXiv2025

Error-controlled non-additive interaction discovery in machine learning models

Machine learning (ML) models are powerful tools for detecting complex patterns within data, yet their "black box" nature limits their interpretability, hindering their use in critical domains like healthcare and finance. To address this challenge, interpretable ML methods have been developed to explain how features influence model predictions. However, these methods often focus on univariate feature importance, overlooking the complex interactions between features that ML models are capable of capturing. Recognizing this limitation, recent efforts have aimed to extend these methods to discover feature interactions, but existing approaches struggle with robustness and error control, especially under data perturbations. In this study, we introduce Diamond, a novel method for trustworthy feature interaction discovery. Diamond uniquely integrates the model-X knockoffs framework to control the false discovery rate (FDR), ensuring that the proportion of falsely discovered interactions remains low. A key innovation in Diamond is its non-additivity distillation procedure, which refines existing interaction importance measures to distill non-additive interaction effects, ensuring that FDR control is maintained. This approach addresses the limitations of off-the-shelf interaction measures, which, when used naively, can lead to inaccurate discoveries. Diamond's applicability spans a wide range of ML models, including deep neural networks, transformer models, tree-based models, and factorization-based models. Our empirical evaluations on both simulated and real datasets across various biomedical studies demonstrate Diamond's utility in enabling more reliable data-driven scientific discoveries. This method represents a significant step forward in the deployment of ML models for scientific innovation and hypothesis generation.

cs.LG↗

arXiv2023

DeepROCK: Error-controlled interaction detection in deep neural networks

The complexity of deep neural networks (DNNs) makes them powerful but also makes them challenging to interpret, hindering their applicability in error-intolerant domains. Existing methods attempt to reason about the internal mechanism of DNNs by identifying feature interactions that influence prediction outcomes. However, such methods typically lack a systematic strategy to prioritize interactions while controlling confidence levels, making them difficult to apply in practice for scientific discovery and hypothesis validation. In this paper, we introduce a method, called DeepROCK, to address this limitation by using knockoffs, which are dummy variables that are designed to mimic the dependence structure of a given set of features while being conditionally independent of the response. Together with a novel DNN architecture involving a pairwise-coupling layer, DeepROCK jointly controls the false discovery rate (FDR) and maximizes statistical power. In addition, we identify a challenge in correctly controlling FDR using off-the-shelf feature interaction importance measures. DeepROCK overcomes this challenge by proposing a calibration procedure applied to existing interaction importance measures to make the FDR under control at a target level. Finally, we validate the effectiveness of DeepROCK through extensive experiments on simulated and real datasets.

cs.LG↗

arXiv2023

Bounding the FDP in competition-based control of the FDR

Competition-based approach to controlling the false discovery rate (FDR) recently rose to prominence when, generalizing it to sequential hypothesis testing, Barber and Candès used it as part of their knockoff-filter. Control of the FDR implies that the, arguably more important, false discovery proportion is only controlled in an average sense. We present TDC-SB and TDC-UB that provide upper prediction bounds on the FDP in the list of discoveries generated when controlling the FDR using competition. Using simulated and real data we show that, overall, our new procedures offer significantly tighter upper bounds than ones obtained using the recently published approach of Katsevich and Ramdas, even when the latter is further improved using the interpolation concept of Goeman et al.

stat.ME↗

arXiv2022

Competition-based control of the false discovery proportion

Recently, Barber and Candès laid the theoretical foundation for a general framework for false discovery rate (FDR) control based on the notion of "knockoffs." A closely related FDR control methodology has long been employed in the analysis of mass spectrometry data, referred to there as "target-decoy competition" (TDC). However, any approach that aims to control the FDR, which is defined as the expected value of the false discovery proportion (FDP), suffers from a problem. Specifically, even when successfully controlling the FDR at level $α$, the FDP in the list of discoveries can significantly exceed $α$. We offer FDP-SD, a new procedure that rigorously controls the FDP in the competition (knockoff / TDC) setup by guaranteeing that the FDP is bounded by $α$ at any desired confidence level. Compared with the just-published general framework of Katsevich and Ramdas, FDP-SD generally delivers more power and often substantially so in simulated as well as real data.

stat.ME↗

arXiv2021

DANCE: Enhancing saliency maps using decoys

Saliency methods can make deep neural network predictions more interpretable by identifying a set of critical features in an input sample, such as pixels that contribute most strongly to a prediction made by an image classifier. Unfortunately, recent evidence suggests that many saliency methods poorly perform, especially in situations where gradients are saturated, inputs contain adversarial perturbations, or predictions rely upon inter-feature dependence. To address these issues, we propose a framework that improves the robustness of saliency methods by following a two-step procedure. First, we introduce a perturbation mechanism that subtly varies the input sample without changing its intermediate representations. Using this approach, we can gather a corpus of perturbed data samples while ensuring that the perturbed and original input samples follow the same distribution. Second, we compute saliency maps for the perturbed samples and propose a new method to aggregate saliency maps. With this design, we offset the gradient saturation influence upon interpretation. From a theoretical perspective, we show the aggregated saliency map could not only capture inter-feature dependence but, more importantly, robustify interpretation against previously described adversarial perturbation methods. Following our theoretical analysis, we present experimental results suggesting that, both qualitatively and quantitatively, our saliency method outperforms existing methods.

cs.LG↗

arXiv2019

Multiple competition-based FDR control for peptide detection

Competition-based FDR control has been commonly used for over a decade in the computational mass spectrometry community (Elias and Gygi, 2007). Recently, the approach has gained significant popularity in other fields after Barber and Candes (2015) laid its theoretical foundation in a more general setting that included the feature selection problem. In both cases, the competition is based on a head-to-head comparison between an observed score and a corresponding decoy / knockoff. Keich and Noble (2017b) recently demonstrated some advantages of using multiple rather than a single decoy when addressing the problem of assigning peptide sequences to observed mass spectra. In this work, we consider a related problem -- detecting peptides based on a collection of mass spectra -- and we develop a new framework for competition-based FDR control using multiple null scores. Within this framework, we offer several methods, all of which are based on a novel procedure that rigorously controls the FDR in the finite sample setting. Using real data to study the peptide detection problem we show that, relative to existing single-decoy methods, our approach can increase the number of discovered peptides by up to 50% at small FDR thresholds.

stat.ME↗

arXiv2019

apricot: Submodular selection for data summarization in Python

We present apricot, an open source Python package for selecting representative subsets from large data sets using submodular optimization. The package implements an efficient greedy selection algorithm that offers strong theoretical guarantees on the quality of the selected set. Two submodular set functions are implemented in apricot: facility location, which is broadly applicable but requires memory quadratic in the number of examples in the data set, and a feature-based function that is less broadly applicable but can scale to millions of examples. Apricot is extremely efficient, using both algorithmic speedups such as the lazy greedy algorithm and code optimizers such as numba. We demonstrate the use of subset selection by training machine learning models to comparable accuracy using either the full data set or a representative subset thereof. This paper presents an explanation of submodular selection, an overview of the features in apricot, and an application to several data sets. The code and tutorial Jupyter notebooks are available at https://github.com/jmschrei/apricot

cs.LG↗

arXiv2018

DeepPINK: reproducible feature selection in deep neural networks

Deep learning has become increasingly popular in both supervised and unsupervised machine learning thanks to its outstanding empirical performance. However, because of their intrinsic complexity, most deep learning methods are largely treated as black box tools with little interpretability. Even though recent attempts have been made to facilitate the interpretability of deep neural networks (DNNs), existing methods are susceptible to noise and lack of robustness. Therefore, scientists are justifiably cautious about the reproducibility of the discoveries, which is often related to the interpretability of the underlying statistical models. In this paper, we describe a method to increase the interpretability and reproducibility of DNNs by incorporating the idea of feature selection with controlled error rate. By designing a new DNN architecture and integrating it with the recently proposed knockoffs framework, we perform feature selection with a controlled error rate, while maintaining high power. This new method, DeepPINK (Deep feature selection using Paired-Input Nonlinear Knockoffs), is applied to both simulated and real data sets to demonstrate its empirical utility.

cs.LG↗

arXiv2012

On the feasibility and utility of exploiting real time database search to improve adaptive peak selection

Rationale: In a shotgun proteomics experiment with data-dependent acquisition, real-time analysis of a precursor scan results in selection of a handful of peaks for subsequent isolation, fragmentation and secondary scanning. This peak selection protocol typically focuses on the most abundant peaks in the precursor scan, while attempting to avoid re-sampling the same m/z values in rapid succession. The protocol does not, however, incorporate analysis of previous fragmentation scans into the peak selection procedure. Methods: In this work, we investigate the feasibility and utility of incorporating analysis of previous fragmentation scans into the peak selection protocol. We demonstrate that real-time identification of fragmentation spectra is feasible in principle, and we investigate, via simulations, several strategies to make use of the resulting peptide identifications during peak selection. Results: Our simulations fail to provide evidence that peptide identifications can provide a large improvement in the total number of peptides identified by a shotgun proteomics experiment. Conclusions: These results are significant because they point out the feasibility of using peptide identifications during peak selection, and because our experiments may provide a starting point for others working in this direction.

q-bio.QM↗

arXiv2006

Metric learning pairwise kernel for graph inference

Much recent work in bioinformatics has focused on the inference of various types of biological networks, representing gene regulation, metabolic processes, protein-protein interactions, etc. A common setting involves inferring network edges in a supervised fashion from a set of high-confidence edges, possibly characterized by multiple, heterogeneous data sets (protein sequence, gene expression, etc.). Here, we distinguish between two modes of inference in this setting: direct inference based upon similarities between nodes joined by an edge, and indirect inference based upon similarities between one pair of nodes and another pair of nodes. We propose a supervised approach for the direct case by translating it into a distance metric learning problem. A relaxation of the resulting convex optimization problem leads to the support vector machine (SVM) algorithm with a particular kernel for pairs, which we call the metric learning pairwise kernel (MLPK). We demonstrate, using several real biological networks, that this direct approach often improves upon the state-of-the-art SVM for indirect inference with the tensor product pairwise kernel.

q-bio.QM↗