Search

Abstract We consider a change-point detection problem for a simple class of Piecewise Deterministic Markov Processes (PDMPs). A continuous-time PDMP is observed in discrete time and through noise, and the aim is to propose a numerical method to accurately detect both the date of the change of dynamics and the new regime after the change. To do so, we state the problem as an optimal stopping problem for a partially observed discrete-time Markov decision process taking values in a continuous state space and provide a discretization of the state space based on quantization to approximate the value function and build a tractable stopping policy. We provide error bounds for the approximation of the value function and numerical simulations to assess the performance of our candidate policy. [ABSTRACT FROM AUTHOR]

Abstract: In this paper, we consider the Integrated Completed Likelihood (ICL) as a useful criterion for estimating the number of changes in the underlying distribution of data, specifically in problems where detecting the precise location of these changes is the main goal. The exact computation of the ICL requires operations (with K the number of segments and n the number of data-points) which is prohibitive in many practical situations with large sequences of data. We describe a framework to estimate the ICL with complexity. Our approach is general in the sense that it can accommodate any given model distribution. We checked the run-time and validity of our approach on simulated data and demonstrate its good performance when analyzing real Next-Generation Sequencing (NGS) data using a negative binomial model. Our method is implemented in the R package postCP and available on the CRAN repository. [Copyright &y& Elsevier]

Transcriptome sequencing (RNA-Seq) yields massive data sets, containing a wealth of information on the expression of a genome. While numerous methods have been developed for the analysis of differential gene expression, little has been attempted for the localization of transcribed regions, that is, segments of DNA that are transcribed and processed to result in a mature messenger RNA. Our understanding of genomes, mostly annotated from biological experiments or computational gene prediction methods, could benefit greatly from re-annotation using the high precision of RNA-Seq. We consider five classes of genome segmentation methods to delineate transcribed regions, including intron/exon boundaries, based on RNA-Seq data. The methods provide different functionality and include both exact and heuristic approaches, using diverse models, such as hidden Markov or Bayesian models, and diverse algorithms, such as dynamic programming or the forward-backward algorithm. We evaluate the methods in a simulation study where RNA-Seq read counts are generated from parametric models as well as by resampling of actual yeast RNA-Seq data. The methods are compared in terms of criteria that include global and local fit to a reference segmentation, Receiver Operator Characteristic (ROC) curves, and coverage of credibility intervals based on posterior change-point distributions. All compared algorithms are implemented in packages available on the Comprehensive R Archive Network (CRAN, ). The data set used in the simulation study is publicly available from the Sequence Read Archive (SRA, ). While the different methods each have pros and cons, our results suggest that the EBS Bayesian approach of Rigaill, Lebarbier, and Robin () performs well in a re-annotation context, as illustrated in the simulation study and in the application to actual yeast RNA-Seq data. This article has supplementary material online. [ABSTRACT FROM AUTHOR]

Designing patient-specific follow-up strategies is key to personalized cancer care. Tools to assist doctors in treatment decisions and scheduling follow-ups based on patient preferences and medical data would be highly beneficial. These tools should incorporate realistic models of disease progression under treatment, multi-objective optimization of treatment strategies, and efficient algorithms to personalize follow-ups by considering patient history. We propose modeling cancer evolution using a Piecewise Deterministic Markov Process, where patients alternate between remission and relapse phases, and control the model via long-term cost function optimization. This considers treatment side effects, visit burden, and quality of life, using noisy blood marker measurements for feedback. Instead of discretizing the problem with a discrete Markov Decision Process, we apply the Partially-Observed Monte-Carlo Planning algorithm to solve the continuous-time, continuous-state problem, leveraging the near-deterministic nature of cancer progression. Our approach, tested on multiple myeloma patient data, outperforms exact solutions of the discrete model and allows greater flexibility in cost function modeling, enabling patient-specific follow-ups. This method can also be adapted to other diseases. [ABSTRACT FROM AUTHOR]

In multiple myeloma, cytogenetic changes are important predictors of patient outcome. In this setting, the most important changes are deletion 17p, del(17p), and translocation of chromosomes 4 and 14, t(4;14), conferring a poor outcome. However, a certain degree of heterogeneity is observed in the survival of these high-risk patients. We hypothesized that other chromosomal changes may impact the outcome. We retrospectively analyzed a large series of 242 patients displaying either t(4;14) (157 patients) or del(17p) (110 patients), 25 patients presenting both abnormalities, using single nucleotide polymorphism array. In patients with t(4;14), del(1p32), del22q, and >30 chromosomal structural changes negatively impacted progression-free survival (PFS). For overall survival (OS), del(13q14), del(1p32), and the number of chromosomal structural changes worsened the prognosis of patients. For patients with del(17p), del6q worsened the prognosis of patients, whereas trisomy 15 and monosomy 14 were found to have a protective effect on PFS. For OS, del(1p32) worsened the prognosis of patients, whereas having >8 numerical changes was found to have a protective effect on survival. This study, which is the largest series of high-risk patients analyzed with the most modern genomic technique, identified 1 main factor negatively impacting survival: del(1p32). [ABSTRACT FROM AUTHOR]

The prognosis of multiple myeloma is mainly dependent upon chromosomal changes. The 2 major abnormalities driving poor outcome are del(17p) and t(4;14). However, the outcome of these high-risk patients is not absolutely uniform, with some patients presenting long survival. We hypothesized that these better outcomes might be related to concomitant "good-risk" chromosomal changes exploring hyperdiploidy. We analyzed a large series of 965 myeloma patients, including 168 patients with t(4;14) and 126 patients with del(17p), using high-throughput single-nucleotide polymorphism arrays after plasma cell sorting. As expected, trisomic chromosomes were highly associated. Using the LASSO model, we found that only chromosome 3, when trisomic, was associated with a longer progression-free survival and that 3 trisomies modulated overall survival (OS) in myeloma patients: trisomies 3 and 5 significantly improved OS, whereas trisomy 21 worsened OS. In patients with t(4;14), trisomies 3 and/or 5 seemed to overcome the poor prognosis. For the first time, using a specific modeling approach, we show that not all trisomies display the same prognostic impact. This finding could be important for routine assessment of prognosis in myeloma, and some high-risk patients with a traditional evaluation could in fact be standard-risk patients. [ABSTRACT FROM AUTHOR]