< Terug naar vorige pagina

Publicatie

varAmpliCNV

Tijdschriftbijdrage - Tijdschriftartikel

Ondertitel:analyzing variance of amplicons to detect CNVs in targeted NGS data
Motivation: Computational identification of copy number variants (CNVs) in sequencing data is a challenging task. Existing CNV-detection methods account for various sources of variation and perform different normalization strategies. However, their applicability and predictions are restricted to specific enrichment protocols. Here, we introduce a novel tool named varAmpliCNV, specifically designed for CNV-detection in amplicon-based targeted resequencing data (Haloplex (TM) enrichment protocol) in the absence of matched controls. VarAmpliCNV utilizes principal component analysis (PCA) and/or metric dimensional scaling (MDS) to control variances of amplicon associated read counts enabling effective detection of CNV signals. Results: Performance of VarAmpliCNV was compared against three existing methods (ConVaDING, ONCOCNV and DECoN) on data of 167 samples run with an aortic aneurysm gene panel (n = 30), including 9 positive control samples. Additionally, we validated the performance on a large deafness gene panel (n = 145) run on 138 samples, containing 4 positive controls. VarAmpliCNV achieved higher sensitivity (100%) and specificity (99.78%) in comparison to competing methods. In addition, unsupervised clustering of CNV segments and visualization plots of amplicons spanning these regions are included as a downstream strategy to filter out false positives. Availability and implementation: he tool is freely available through galaxy toolshed and at:https://hub.docker.com/r/cmgantwerpen/varamplicnv.Supplementary Data File S1:https://tinyurl.com/2yzswyhh;Supplementary DataFile S2:https://tinyurl.com/ycyf2fb4. Contact: maaike.alaerts @uantwerpen.be or geert.vandeweyer@uantwerpen.be Supplementary information: Supplementary dataare available atBioinformatics online.
Tijdschrift: Bioinformatics
ISSN: 1367-4803
Volume: 31
Pagina's: 1 - 8
Jaar van publicatie:2023
Trefwoorden:A1 Journal article
Toegankelijkheid:Open