Increasing diagnostic yield of fast whole genome diagnostics for new borns on the Neonatal Intensive Care Unit (NICU)

Supervisors

Pieter Neerincx and Freerk van Dijk

Introduction

Next Generation Sequencing (NGS) has become a very important tool for geneticists. At the UMCG we recently enrolled in project "5 Genes per Minute (5GpM)" to move this high throughput technology from the research department into the clinic. In 5 GpM we aim to diagnose genetic disease in severely affected new borns on our Neonatal Intensive Care Unit (NICU) by means of whole genome screening. 5 Genes per Minute analysis can be requested for patients when all classic diagnostics failed and speed is essential. We aim for from blood sample to diagnosis in < 48 hours.

Our current bioinformatics analysis only considers single nucleotide variants (SNVs) and short insertion/deletions (InDels). So far we have analysed 10 patients and managed to pin point the causal genetic variant in one. There may be various reasons why we have no diagnosis in the other 9:

• The symptoms are not caused by a genetic disease: something else is going on.
• The genetic variant is part of a region that cannot be reliably sequenced.
• The genetic variant was sequenced, but is not yet called by our bioinformatics pipeline.
• The genetic variant was sequenced and called, but we cannot interpret the effect of the variant correctly.

The largest amount of genetic variants currently missed are either Copy Number Variants (CNVs) and medium sized InDels: larger than the NGS read length, but shorter than what can be detected with classic diagnostics like array CGH (aCGH). An additional complication for the current short InDel calls is that only the variant is detected, but not the genotype (heterozygous vs. homozygous). Large insertions are expected less frequently, but not less important. Calls for this type of variants need to be included in the analysis too. Finally the calls (of any type) are not yet optimized for non-autosomal chormosomes like the sex chromosomes and the mitochondrial chromosome.

Projects 1 through 5 - Improved variant calling

In these student research projects we will choose a type of chromosome or type of genetic variant for which we will evaluate 2-3 new state of the art algorithms/tools. We'll add the best performing one to our analysis pipeline, re-analyse our unsolved cases and test if we can increase the diagnostic yield for this group of patients.