Niko Popitsch

NGC: lossless and lossy compression of aligned high-throughput sequencing data

A major challenge of current high-throughput sequencing experiments is not only the generation of the sequencing data itself but also their processing, storage and transmission. The enormous size of these data motivates the development of data compression algorithms usable for the implementation of the various storage policies that are applied to the produced intermediate and final result files. In this article, we present NGC, a tool for the compression of mapped short read data stored in the wide-spread SAM format. NGC enables lossless and lossy compression and introduces the following two novel ideas: first, we present a way to reduce the number of required code words by exploiting common features of reads mapped to the same genomic positions; second, we present a highly configurable way for the quantization of per-base quality values, which takes their influence on downstream analyses into account. NGC, evaluated with several real-world data sets, saves 33–66% of disc space using lossless and up to 98% disc space using lossy compression. By applying two popular variant and genotype prediction tools to the decompressed data, we could show that the lossy compression modes preserve >99% of all called variants while outperforming comparable methods in some configurations.

Revisiting the coding potential of the E. coli genome through Hfq co-immunoprecipitation

Hfq is a global regulator of gene expression in bacteria undergoing adaptation to changing environmental conditions. Its major function is to promote RNA-RNA interactions between regulatory small RNAs (sRNAs) and their target mRNAs. Previously, we demonstrated that Hfq binds many antisense RNAs (asRNAs) in vitro and hypothesized that Hfq may play a role in regulating gene expression via asRNAs. To investigate the E. coli Hfq-binding transcriptome in more detail, we co-immunoprecipitated and deep-sequenced RNAs bound to Hfq in vivo. We detected many new Hfq-binding sRNAs and observed that almost 300 mRNAs bind to Hfq. Among these, several are known to be sRNA targets. We identified 25 novel RNAs, which are transcribed from within protein coding regions and named them intragenic RNAs (intraRNAs). Furthermore, 67 asRNAs were co-immunoprecipitated with Hfq, demonstrating that Hfq binds antisense transcripts in vivo. Northern blot analyses confirmed the deep sequencing results and demonstrated that many of the novel Hfq-binding RNAs identified are regulated by Hfq.

The Borrelia burgdorferi RelA/SpoT Homolog and Stringent Response Regulate Survival in the Tick Vector and Global Gene Expression during Starvation

As the Lyme disease bacterium Borrelia burgdorferi traverses its enzootic cycle, alternating between a tick vector and a vertebrate host, the spirochete must adapt and persist in the tick midgut under prolonged nutrient stress between blood meals. In this study, we examined the role of the stringent response in tick persistence and in regulation of gene expression during nutrient limitation. Nutritionally starving B. burgdorferi in vitro increased the levels of guanosine tetraphosphate (ppGpp) and guanosine pentaphosphate (pppGpp), collectively referred to as (p)ppGpp, products of the bifunctional synthetase/hydrolase RelBbu (RelA/SpoT homolog). Conversely, returning B. burgdorferi to a nutrient-rich medium decreased (p)ppGpp levels. B. burgdorferi survival in ticks between the larval and nymph blood meals, and during starvation in vitro, was dependent on RelBbu. Furthermore, normal morphological conversion from a flat-wave shape to a condensed round body (RB) form during starvation was dependent on RelBbu; rel Bbu mutants more frequently formed RBs, but their membranes were compromised. By differential RNA sequencing analyses, we found that RelBbu regulates an extensive transcriptome, both dependent and independent of nutrient stress. The RelBbu regulon includes the glp operon, which is important for glycerol utilization and persistence in the tick, virulence factors and the late phage operon of the 32-kb circular plasmid (cp32) family. In summary, our data suggest that RelBbu globally modulates transcription in response to nutrient stress by increasing (p)ppGpp levels to facilitate B. burgdorferi persistence in the tick.

Mutations in PIGY: expanding the phenotype of inherited glycosylphosphatidylinositol deficiencies

Glycosylphosphatidylinositol (GPI)-anchored proteins are ubiquitously expressed in the human body and are important for various functions at the cell surface. Mutations in many GPI biosynthesis genes have been described to date in patients with multi-system disease and together these constitute a subtype of congenital disorders of glycosylation. We used whole exome sequencing in two families to investigate the genetic basis of disease and used RNA and cellular studies to investigate the functional consequences of sequence variants in the PIGY gene. Two families with different phenotypes had homozygous recessive sequence variants in the GPI biosynthesis gene PIGY. Two sisters with c.137T>C (p.Leu46Pro) PIGY variants had multi-system disease including dysmorphism, seizures, severe developmental delay, cataracts and early death. There were significantly reduced levels of GPI-anchored proteins (CD55 and CD59) on the surface of patient-derived skin fibroblasts (∼20–50% compared with controls). In a second, consanguineous family, two siblings had moderate development delay and microcephaly. A homozygous PIGY promoter variant (c.-540G>A) was detected within a 7.7 Mb region of autozygosity. This variant was predicted to disrupt a SP1 consensus binding site and was shown to be associated with reduced gene expression. Mutations in PIGY can occur in coding and non-coding regions of the gene and cause variable phenotypes. This article contributes to understanding of the range of disease phenotypes and disease genes associated with deficiencies of the GPI-anchor biosynthesis pathway and also serves to highlight the potential importance of analysing variants detected in 5′-UTR regions despite their typically low coverage in exome data.

In vivo expression technology and 5′ end mapping of the Borrelia burgdorferi transcriptome identify novel RNAs expressed during mammalian infection

Borrelia burgdorferi, the bacterial pathogen responsible for Lyme disease, modulates its gene expression profile in response to the environments encountered throughout its tick-mammal infectious cycle. To begin to characterize the B. burgdorferi transcriptome during murine infection, we previously employed an in vivo expression technology-based approach (BbIVET). This identified 233 putative promoters, many of which mapped to un-annotated regions of the complex, segmented genome. Herein, we globally identify the 5′ end transcriptome of B. burgdorferi grown in culture as a means to validate non-ORF associated promoters discovered through BbIVET. We demonstrate that 119 BbIVET promoters are associated with transcription start sites (TSSs) and validate novel RNA transcripts using Northern blots and luciferase promoter fusions. Strikingly, 49% of BbIVET promoters were not found to associate with TSSs. This finding suggests that these sequences may be primarily active in the mammalian host. Furthermore, characterization of the 6042 B. burgdorferi TSSs reveals a variety of RNAs including numerous antisense and intragenic transcripts, leaderless RNAs, long untranslated regions and a unique nucleotide frequency for initiating intragenic transcription. Collectively, this is the first comprehensive map of TSSs in B. burgdorferi and characterization of previously un-annotated RNA transcripts expressed by the spirochete during murine infection.

Germline recessive mutations in PI4KA are associated with perisylvian polymicrogyria, cerebellar hypoplasia and arthrogryposis

Polymicrogyria (PMG) is a structural brain abnormality involving the cerebral cortex that results from impaired neuronal migration and although several genes have been implicated, many cases remain unsolved. In this study, exome sequencing in a family where three fetuses had all been diagnosed with PMG and cerebellar hypoplasia allowed us to identify regions of the genome for which both chromosomes were shared identical-by-descent, reducing the search space for causative variants to 8.6% of the genome. In these regions, the only plausibly pathogenic mutations were compound heterozygous variants in PI4KA, which Sanger sequencing confirmed segregated consistent with autosomal recessive inheritance. The paternally transmitted variant predicted a premature stop mutation (c.2386C>T; p.R796X), whereas the maternally transmitted variant predicted a missense substitution (c.5560G>A; p.D1854N) at a conserved residue within the catalytic domain. Functional studies using expressed wild-type or mutant PI4KA enzyme confirmed the importance of p.D1854 for kinase activity. Our results emphasize the importance of phosphoinositide signalling in early brain development.

Analysis of exome data for 4293 trios suggests GPI-anchor biogenesis defects are a rare cause of developmental disorders

Over 150 different proteins attach to the plasma membrane using glycosylphosphatidylinositol (GPI) anchors. Mutations in 18 genes that encode components of GPI-anchor biogenesis result in a phenotypic spectrum that includes learning disability, epilepsy, microcephaly, congenital malformations and mild dysmorphic features. To determine the incidence of GPI-anchor defects, we analysed the exome data from 4293 parent–child trios recruited to the Deciphering Developmental Disorders (DDD) study. All probands recruited had a neurodevelopmental disorder. We searched for variants in 31 genes linked to GPI-anchor biogenesis and detected rare biallelic variants in PGAP3, PIGN, PIGT (n=2), PIGO and PIGL, providing a likely diagnosis for six families. In five families, the variants were in a compound heterozygous configuration while in a consanguineous Afghani kindred, a homozygous c.709G>C; p.(E237Q) variant in PIGT was identified within 10–12 Mb of autozygosity. Validation and segregation analysis was performed using Sanger sequencing. Across the six families, five siblings were available for testing and in all cases variants co-segregated consistent with them being causative. In four families, abnormal alkaline phosphatase results were observed in the direction expected. FACS analysis of knockout HEK293 cells that had been transfected with wild-type or mutant cDNA constructs demonstrated that the variants in PIGN, PIGT and PIGO all led to reduced activity. Splicing assays, performed using leucocyte RNA, showed that a c.336-2A>G variant in PIGL resulted in exon skipping and p.D113fs*2. Our results strengthen recently reported disease associations, suggest that defective GPI-anchor biogenesis may explain ~0.15% of individuals with developmental disorders and highlight the benefits of data sharing.

Spectrum of gene mutations detected by next generation exome sequencing in brain metastases of lung adenocarcinoma

BACKGROUND Brain metastases (BM) are a life-threatening complication. We aimed to analyse gene mutations in lung adenocarcinoma BM. METHODS We performed next generation sequencing (NGS) of a pre-defined set of 48 cancer-related genes in a cohort of 76 neurosurgical lung adenocarcinoma BM specimens using a cancer specific gene panel on the MiSeq platform (Illumina, San Diego, CA). NGS results were statistically correlated to patient characteristics. Data on ALK, ROS1, MET and FGFR1 gene status assessed by FISH were available from previous studies in the majority of patients. RESULTS Twenty-nine (60.4%) of the 48 investigated cancer-related genes were mutated in at least one BM sample and 64 (84.2%) of the 76 BM samples carried at least one mutated gene. The number of mutated genes per sample ranged from 0 to 9 (median 2). The most commonly mutated genes were TP53, KRAS and CDKN2A, which were affected in 35/76 (46.1%), 29/76 (38.2%) and 17/76 (22.4%) samples, respectively. Other potentially druggable alterations included EGFR mutations (3/76, 3.9% of samples), PIK3CA mutation (2/76, 2.6%), BRAF mutation (1/76, 1.3%) and SMO mutation (1/76, 1.3%). Presence of KRAS mutations was associated with positive smoking history (p=0.015, Chi square test) and presence of EGFR mutation correlated with unfavourable overall survival time from BM diagnosis (p=0.019, log rank test). CONCLUSIONS Deleterious gene mutations, some of them with potential therapeutic implications, are found in a high fraction of lung adenocarcinoma BM.

Targeted Next-Generation Sequencing of Plasma DNA from Cancer Patients: Factors Influencing Consistency with Tumour DNA and Prospective Investigation of Its Utility for Diagnosis

Use of circulating tumour DNA (ctDNA) as a liquid biopsy has been proposed for potential identification and monitoring of solid tumours. We investigate a next-generation sequencing approach for mutation detection in ctDNA in two related studies using a targeted panel. The first study was retrospective, using blood samples taken from melanoma patients at diverse timepoints before or after treatment, aiming to evaluate correlation between mutations identified in biopsy and ctDNA, and to acquire a first impression of influencing factors. We found good concordance between ctDNA and tumour mutations of melanoma patients when blood samples were collected within one year of biopsy or before treatment. In contrast, when ctDNA was sequenced after targeted treatment in melanoma, mutations were no longer found in 9 out of 10 patients, suggesting the method might be useful for detecting treatment response. Building on these findings, we focused the second study on ctDNA obtained before biopsy in lung patients, i.e. when a tentative diagnosis of lung cancer had been made, but no treatment had started. The main objective of this prospective study was to evaluate use of ctDNA in diagnosis, investigating the concordance of biopsy and ctDNA-derived mutation detection. Here we also found positive correlation between diagnostic lung biopsy results and pre-biopsy ctDNA sequencing, providing support for using ctDNA as a cost-effective, non-invasive solution when the tumour is inaccessible or when biopsy poses significant risk to the patient.

Clinical whole-genome sequencing from routine formalin-fixed, paraffin-embedded specimens: pilot study for the 100,000 Genomes Project

PurposeFresh-frozen (FF) tissue is the optimal source of DNA for whole-genome sequencing (WGS) of cancer patients. However, it is not always available, limiting the widespread application of WGS in clinical practice. We explored the viability of using formalin-fixed, paraffin-embedded (FFPE) tissues, available routinely for cancer patients, as a source of DNA for clinical WGS.MethodsWe conducted a prospective study using DNAs from matched FF, FFPE, and peripheral blood germ-line specimens collected from 52 cancer patients (156 samples) following routine diagnostic protocols. We compared somatic variants detected in FFPE and matching FF samples.ResultsWe found the single-nucleotide variant agreement reached 71% across the genome and somatic copy-number alterations (CNAs) detection from FFPE samples was suboptimal (0.44 median correlation with FF) due to nonuniform coverage. CNA detection was improved significantly with lower reverse crosslinking temperature in FFPE DNA extraction (80 °C or 65 °C depending on the methods). Our final data showed somatic variant detection from FFPE for clinical decision making is possible. We detected 98% of clinically actionable variants (including 30/31 CNAs).ConclusionWe present the first prospective WGS study of cancer patients using FFPE specimens collected in a routine clinical environment proving WGS can be applied in the clinic.