VINYL: Variant prIoritizatioN bY survivaL analysis

Abstract

Motivation Clinical applications of genome re-sequencing technologies typically generate large amounts of data that need to be carefully annotated and interpreted to identify genetic variants potentially associated with pathological conditions. In this context, accurate and reproducible methods for the functional annotation and prioritization of genetic variants are of fundamental importance. Results In this article, we present VINYL, a flexible and fully automated system for the functional annotation and prioritization of genetic variants. Extensive analyses of both real and simulated datasets suggest that VINYL can identify clinically relevant genetic variants in a more accurate manner compared to equivalent state of the art methods, allowing a more rapid and effective prioritization of genetic variants in different experimental settings. As such we believe that VINYL can establish itself as a valuable tool to assist healthcare operators and researchers in clinical genomics investigations. Availability and implementation VINYL is available at http://beaconlab.it/VINYL and https://github.com/matteo14c/VINYL. Supplementary information Supplementary data are available at Bioinformatics online.

Laniakea@ReCaS: exploring the potential of customisable Galaxy on-demand instances as a cloud-based service

“…To mitigate, at least in part, some of these issues and improve the reproducibility of variant prioritization in clinical studies, our research group from the University of Milan and the CNR Institute of Biomembranes, Bioenergetics and Molecular Biotechnologies (CNR-IBIOM) have recently developed VINYL (Variant prIoritizatioN bY survivaL analysis). This novel software suite integrates an innovative method for variant prioritization, along with a highly curated collection of databases and resources for the annotation of human genetic variants [ 51 ]. The main advantage of VINYL over other existing methods is that our tool can evaluate different scoring systems and metrics for the prioritization of genetic variants and provides a fully automated procedure to derive optimal criteria for the identification of genetic variants of potential clinical relevance.…”

“…The system is flexible and allows the design of custom scoring schemes based on personalized functional annotations and can be adapted/optimized to different use cases and scenarios. Notably, extensive comparisons with equivalent state of the art methods demonstrated that VINYL could detect different types of genetic variants associated with pathological conditions and achieve higher levels of sensitivity and specificity than equivalent state of the art methods [ 51 ].…”

“…All the required software and wrappers were also made available through Github to facilitate the installation and configuration of VINYL also on novel Galaxy instances [ 52 ]. A paper describing the method and the Galaxy instance was accepted for publication last December [ 51 ]. The VINYL Galaxy [ 53 ] service is officially up since the acceptance of the paper with no downtime.…”

Survival Time Prediction of Breast Cancer Patients Using Feature Selection Algorithm Crystall

“…Firstly, we evaluated various classification-based feature selection algorithms on the binary classification problem of whether a patient lived longer than five years or not. The status of five-year survival was widely investigated in the conventional survival analysis studies [39,40], and a patient was generally considered as being "cancer free" if this patient lived for five years and longer [41][42][43][44]. Each feature selection algorithm was applied separately on the dataset and the subset of chosen features was used to build the classification model using one of the five classifiers with the 10-fold cross validation strategy.…”

“…Survival time prediction is an important question for a patient with a lethal disease. This is different from the conventional survival analysis which tries to calculate the percentage of alive patients within a cohort on a time point [39,40]. Nie et al utilized the 3-dimensional convolution neural network (CNN) to extract features for an SVM model and predicted with an accuracy 89.9% whether a patient had a long or short overall survival time [88].…”