mobidic/BARMEN

BenchmARking from Mobidic for Exome sequencing at Nucleotide scale

BARMEN welcome you !

• BARMEN - The BenchmARking from Mobidic for Exome sequencing at Nucleotide scale
  • Overview
    • Main Workflow
    • Cite BARMEN
    • Input
    • Output
  • Installation
    • Requirements
  • BARMEN command
    • In silico design analysis
    • ClinVar pathogenic variant coverage
    • Gene Panel coverage
    • RefSeq Coding regions coverage

---

The BenchmARking from Mobidic for Exome sequencing at Nucleotide scale aims to compare recent Whole Exome capture library performance with several items :

• In silico design analysis
• Coverage of ClinVar pathogenic variants
• Coverage of gene panel sequencing routine diagnosis
• Coverage of RefSeq Coding regions

Capture library benchmark include :

• Nimblegen MedExome
• Agilent SureSelect V7
• Agilent Clinical Research Exome V2

Overview

Main workflow

For performance benchmark, we use the capture targets bed (and not the empirical or covered bed).

• MedExome_hg19_capture_targets.bed : MedExome capture target intervals converted to UCSC human genome build hg19 using the NCBI Genome Remapping Service.

• Agilent for SureSelect Clinical Research Exome V7 and SureSelect Human All Exon V7 [design ID]_Regions.bed : This BED file contains a single track of the target regions of interest that SureDesign used to select the probes. You can use this track to see the exact regions that the program was attempting to cover when selecting the probes.

Selection of samples from our dataset :

• We have 6 datasets : SSV7 (2x75 enzymatic, 2x150 mecanic) CREV2 (2x75 enzymatic, 2x150 mecanic, 2x150 enzymatic) MedExome (2x150 meca)

• We select one sample from each dataset based on a kd-tree approach (Packer JS et al., Clamms, Bioinformatics). For all sample, we compute the 6 closest neighbough (knns) based on Picard metrics GCDROPOUT, ATDROPOUT, MEANINSERTSIZE, ONBAITVSSELECTED, PCTPFUQREADS, PCTTARGETBASES10X, and PCTTARGETBASES50X. Then we select the closest sample of each dataset to compare each library performance with less batch effects.

Cite BARMEN

Pegeot et al., BARMEN : Benchmarking from Mobidic for Whole Exome sequencing at Nucleotide scale (2019).

Input

Output

Installation

To download scripts of BARMEN, please use git to download the most recent development tree. Currently, the tree is hosted on github, and can be obtained via:

$ git clone https://github.com/mobidic/BARMEN.git

Requirements

• Linux or macOS
• bedtools v2.27.1
• picard

BARMEN command

In silico design analysis

Raw Metrics

ClinVar pathogenic variant coverage

Gene Panel coverage

RefSeq Coding regions coverage

---

Montpellier Bioinformatique pour le Diagnostique Clinique (MoBiDiC)

CHU de Montpellier

France

Visit our website

---