Assessment of targeted next generation sequencing as a tool for the diagnosis of Charcot-Marie-Tooth disease and hereditary motor neuropathy

Charcot-Marie-Tooth disease (CMT) is characterized by broad genetic heterogeneity with more than 50 known disease-associated genes. Mutations in some of these genes can cause a pure motor form of hereditary motor neuropathy (HMN), the genetics of which are poorly characterized. We have designed a panel comprising 56 genes associated with CMT/HMN. We validated this diagnostic tool by first testing 11 patients with pathological mutations. Then a cohort of 33 affected subjects was selected for this study. The DNAJB2 c.352+1G>A mutation was detected in 2 cases; novel changes and/or variants with low frequency (<1%) were found in 12 cases. There were no candidate variants in 18 cases, and amplification failed for one sample The DNAJB2 c.352+1G>A mutation was also detected in three additional families. Haplotype analysis showed that all of the patients from these five families share the same haplotype; therefore, the DNAJB2 c.352+1G>A mutation may be a founder event. Our gene panel allowed us to perform a very rapid and cost-effective screening of genes involved in CMT/HMN. Our diagnostic strategy proved to be robust in terms of both coverage and read depth for all of the genes and patients samples. These findings demonstrate the difficulty to achieve a definitive molecular diagnosis because of the complexity of interpreting new variants and the genetic heterogeneity that is associated with these neuropathies.

• Primary analysis carried out from DNAnexus software. Raw reads were generated from the standard Illumina pipeline with default options.
• The original mappings are just the reads mapped to the genome via BWA mem. The refined mappings are the original mappings which have been put through the GATK IndelRealigner and the GATK BaseRecalibrator. There was not removing duplicates because Haloplex reads are expected to have identical start locations due to the way the procedure works. For each sample, we performed a quality evaluation from SAMStat.
• These are usually best for any downstream analysis and are what we use when we call mutations. VARIANT allowed us to get several statistical indicators about variant calling.
• Annotation variant included minor allele frequency (MAF), obtained from dbSNP, 1000 Genomes project and CIBERER Spanish Variant Server to help on the selection of new variants not reported in healthy population. SIFT and Polyphen damage scores were computed to predict the putative impact of the discovered variants over the protein structure and functionality.
• Sequence data have been deposited in SRA repository under the accession number SRP061110.

• A.2.1. Sequence Processing
• A.2.2. Mapping
• A.2.3. Variant Calling

• To evaluate the sequencing coverage we used BAM files to get coverage indicators from GATK and using the capture kit BED file.
• Clustering and principal components methods were performed to explore the coverage data for all samples. Boxplots, scatterplots and statistics were used to describe coverage by gene and sample. Bar graphs described mean coverage for each region and gene.

Results are shown by several units: region, gene and sample. There are tables including main statistical indicators and plots representing all information for each element (by clicking the graph, you can enlarge this representation).

Detailed statistics by region and run
All coverage data

• A. Relationship between size region and its coverage
• B. Are there coverage differences by run?