The human gene damage index as a gene-level approach to prioritizing exome variants

The protein-coding exome of a patient with a monogenic disease contains about 20,000 variants, only one or two of which are disease causing. We found that 58% of rare variants in the protein-coding exome of the general population are located in only 2% of the genes. Prompted by this observation, we aimed to develop a gene-level approach for predicting whether a given human protein-coding gene is likely to harbor disease-causing mutations. To this end, we derived the gene damage index (GDI): a genome-wide, gene-level metric of the mutational damage that has accumulated in the general population. We found that the GDI was correlated with selective evolutionary pressure, protein complexity, coding sequence length, and the number of paralogs. We compared GDI with the leading gene-level approaches, genic intolerance, and de novo excess, and demonstrated that GDI performed best for the detection of false positives (i.e., removing exome variants in genes irrelevant to disease), whereas genic intolerance and de novo excess performed better for the detection of true positives (i.e., assessing de novo mutations in genes likely to be disease causing). The GDI server, data, and software are freely available to noncommercial users from lab.rockefeller.edu/casanova/GDI.Germ-line mutations can contribute to the long-term adaptation of humans, but at the expense of causing a large number of genetic diseases (1). The advent of next-generation sequencing (NGS)-based approaches, including whole-exome sequencing (WES), whole-genome sequencing (WGS), and RNA-Seq, has facilitated the large-scale detection of gene variants at both the individual and population levels (2–6). In patients suffering from a monogenic disease, at most two variants are disease causing [true positives (TP)], and the other 20,000 or so protein-coding exome variants are false positives (FP; type I error). Several variant-level metrics predicting the biochemical impact of DNA mutations (7–9) can be used to prioritize candidate variants for a phenotype of interest (10, 11). Gene-level metrics aim to prioritize the genes themselves, providing information that can be used for the further prioritization of variants. There are currently fewer gene-level than variant-level computational methods. They provide complementary information, as it is best to predict the impact of a variant by also taking into account population genetics data for its locus. Current gene-level methods include genic intolerance, as measured by the residual variation intolerance score (RVIS) (12) and de novo excess (DNE) (13). These metrics are particularly useful for determining whether a given gene (and, by inference, its variants) is a plausible candidate for involvement in a particular genetic disease (i.e., for the selection of a short list of candidate genes and variants, which include the TPs). However, owing to the large number and diversity of variants, the selection of a single candidate gene from the NGS data for a given patient with a specific disease remains challenging.We reasoned that genes frequently mutated in healthy populations would be unlikely to cause inherited and rare diseases, but would probably make a disproportionate contribution to the variant calls observed in any given patient. Conversely, mutations in genes that are never or only rarely mutated under normal circumstances are more likely to be disease-causing. Leading gene-level strategies are based on selective pressure (12) and de novo mutation rate estimates (13). These methods are tailored to detect genes likely to harbor TPs. However, these methods do not directly calculate quantitatively the mutational load for human genes in the general (i.e., “healthy”) population or the frequencies of mutant alleles. These methods may, therefore, not be optimal for filtering out highly mutated genes, which are likely to harbor many FPs. Moreover, there has been no formal comparison of the power of these gene-level methods and their combinations for maximizing the discovery of FPs and TPs by NGS. We therefore aimed to generate a robust metric of the cumulative mutational damage to each human protein-coding gene, to make it easier to distinguish the FP variants harbored by highly damaged genes (e.g., under relaxed constraint or positive selection) from potential candidate genes and variants, including the TPs. By damaged genes, we refer to genes displaying many nonsynonymous mutations, which are not necessarily damaging biochemically or evolutionarily. We developed the gene damage index (GDI), which defines, in silico, the mutational damage accumulated by each protein-coding human gene in the general population, and reflecting the combined influences of drifts and selections. We then tested this approach with the WES data for 84 patients in our in-house database, each of these patients having a known primary immunodeficiency (PID). Finally, we used receiver operating characteristic (ROC) curves for formal comparisons of performance between GDI and the existing gene-level RVIS and DNE approaches, and to assess the power of the gene-level methods for detecting enrichment in de novo mutations in cases versus controls. We also tested whether these methods could act in synergy to filter out FPs and select TPs.     

AcrB drug-binding pocket substitution confers clinically relevant resistance and altered substrate specificity

The incidence of multidrug-resistant bacterial infections is increasing globally and the need to understand the underlying mechanisms is paramount to discover new therapeutics. The efflux pumps of Gram-negative bacteria have a broad substrate range and transport antibiotics out of the bacterium, conferring intrinsic multidrug resistance (MDR). The genomes of pre- and posttherapy MDR clinical isolates of Salmonella Typhimurium from a patient that failed antibacterial therapy and died were sequenced. In the posttherapy isolate we identified a novel G288D substitution in AcrB, the resistance-nodulation division transporter in the AcrAB-TolC tripartite MDR efflux pump system. Computational structural analysis suggested that G288D in AcrB heavily affects the structure, dynamics, and hydration properties of the distal binding pocket altering specificity for antibacterial drugs. Consistent with this hypothesis, recreation of the mutation in standard Escherichia coli and Salmonella strains showed that G288D AcrB altered substrate specificity, conferring decreased susceptibility to the fluoroquinolone antibiotic ciprofloxacin by increased efflux. At the same time, the substitution increased susceptibility to other drugs by decreased efflux. Information about drug transport is vital for the discovery of new antibacterials; the finding that one amino acid change can cause resistance to some drugs, while conferring increased susceptibility to others, could provide a basis for new drug development and treatment strategies.The incidence of multidrug-resistant (MDR, also used herein for “multidrug resistance”) bacterial infections is increasing, and the 2013 World Economic Forum Global Risks report listed antibiotic-resistant bacteria as one of the greatest threats to human health (1). Resistance-nodulation division (RND) efflux pumps of Gram-negative bacteria confer intrinsic and acquired MDR in clinically relevant infections by exporting antibiotics out of the bacterial cell, allowing bacteria to survive at increased drug concentrations.To date the best-characterized efflux pump is AcrAB-TolC of Escherichia coli, composed of the inner membrane RND antiporter AcrB that functions in a tripartite assembly with a periplasmic adaptor protein, AcrA, and the outer membrane channel, TolC. The AcrB pump is an asymmetric homotrimer whose monomers undergo a functional rotation through three states: access, binding, and extrusion (labeled A, B, and E, respectively) to pump substrates from the periplasm, or outer leaflet of the inner membrane, to outside of the cell (2, 3). Drug binding within AcrB is complex. The distal binding pocket has a phenylalanine-rich region that binds low-molecular-weight drugs and a proximal binding pocket and vestibule bind larger compounds such as erythromycin (4, 5).To date, MDR in clinically relevant infections mediated by RND efflux pumps, including AcrB, has been documented to be due to overexpression of the efflux pump and concomitant increased efflux of antibiotics from the bacterial cell (6). Overproduction of efflux pumps can occur via four mechanisms: (i) mutation of the local repressor gene (7, 8), (ii) mutation in a global regulatory gene (9, 10), (iii) mutation of the promoter region of the efflux pump gene (11), or (iv) insertion elements upstream of the transporter gene (12, 13).This study focused on elucidating the mechanism of resistance in a unique set of clinical isolates collected over the course of a complex Salmonella infection (Fig. S1) (14, 15). The isolates were obtained from a 52-y-old male patient admitted for repair of a leaking abdominal aortic aneurysm graft. Salmonella Typhimurium (L3) was isolated before ciprofloxacin treatment and was susceptible to ampicillin, sulphonamide, trimethoprim, cefuroxime, chloramphenicol, gentamicin, and ciprofloxacin. Over the course of the infection the patient received i.v. ciprofloxacin, oral ciprofloxacin, i.v. ceftazidime, and i.v. aztreonam. Isolates were taken throughout infection, and during treatment MDR strains were isolated. L18, the last MDR strain to be isolated, was from wound drainage fluid when the patient had received no antibiotics for 2 wks. The patient died soon afterward with the infection unresolved. Compared with the pretherapy isolate L3, the MDR posttherapy isolate, L18, was less susceptible to numerous agents, including ciprofloxacin and β-lactams, and accumulated less ciprofloxacin and Hoechst dye (16). This set of isolates has provided a unique opportunity to investigate in vivo evolution of MDR in response to clinically validated courses of antimicrobial treatment.Here, we report that whole genome sequencing revealed a new mechanism of clinically significant MDR selected during therapy: substitution in a transporter protein, which altered the specificity of the efflux pump for antibacterial drugs. The structural impact of the substitution was investigated by molecular dynamics (MD) simulations.     

Novel recurrently mutated genes in African American colon cancers

We used whole-exome and targeted sequencing to characterize somatic mutations in 103 colorectal cancers (CRC) from African Americans, identifying 20 new genes as significantly mutated in CRC. Resequencing 129 Caucasian derived CRCs confirmed a 15-gene set as a preferential target for mutations in African American CRCs. Two predominant genes, ephrin type A receptor 6 (EPHA6) and folliculin (FLCN), with mutations exclusive to African American CRCs, are by genetic and biological criteria highly likely African American CRC driver genes. These previously unsuspected differences in the mutational landscapes of CRCs arising among individuals of different ethnicities have potential to impact on broader disparities in cancer behaviors.Colorectal cancer (CRC) is a leading cause of cancer mortality world-wide. CRC incidence and mortality rates are both increased in African Americans (AA) compared with Caucasians Americans (1–3). Although several factors likely play a role, the contribution of potential differences in tumor genetics to this disparity have yet to be fully explored (1, 3). In particular, AA CRCs were notably underrepresented in the four major published CRC sequencing studies (4–7), accounting for only two annotated AA cases of the 333 total CRCs studied (4–7). Accordingly, we initiated this study to compare the mutational landscapes of CRCs from AA individuals versus Caucasians.     

Whole-genome sequencing is more powerful than whole-exome sequencing for detecting exome variants

We compared whole-exome sequencing (WES) and whole-genome sequencing (WGS) in six unrelated individuals. In the regions targeted by WES capture (81.5% of the consensus coding genome), the mean numbers of single-nucleotide variants (SNVs) and small insertions/deletions (indels) detected per sample were 84,192 and 13,325, respectively, for WES, and 84,968 and 12,702, respectively, for WGS. For both SNVs and indels, the distributions of coverage depth, genotype quality, and minor read ratio were more uniform for WGS than for WES. After filtering, a mean of 74,398 (95.3%) high-quality (HQ) SNVs and 9,033 (70.6%) HQ indels were called by both platforms. A mean of 105 coding HQ SNVs and 32 indels was identified exclusively by WES whereas 692 HQ SNVs and 105 indels were identified exclusively by WGS. We Sanger-sequenced a random selection of these exclusive variants. For SNVs, the proportion of false-positive variants was higher for WES (78%) than for WGS (17%). The estimated mean number of real coding SNVs (656 variants, ∼3% of all coding HQ SNVs) identified by WGS and missed by WES was greater than the number of SNVs identified by WES and missed by WGS (26 variants). For indels, the proportions of false-positive variants were similar for WES (44%) and WGS (46%). Finally, WES was not reliable for the detection of copy-number variations, almost all of which extended beyond the targeted regions. Although currently more expensive, WGS is more powerful than WES for detecting potential disease-causing mutations within WES regions, particularly those due to SNVs.Whole-exome sequencing (WES) is routinely used and is gradually being optimized for the detection of rare and common genetic variants in humans (1–8). However, whole-genome sequencing (WGS) is becoming increasingly attractive as an alternative, due to its broader coverage and decreasing cost (9–11). It remains difficult to interpret variants lying outside the protein-coding regions of the genome. Diagnostic and research laboratories, whether public or private, therefore tend to search for coding variants, most of which can be detected by WES, first. Such variants can also be detected by WGS, and several studies previously compared WES and WGS for different types of variations and/or in different contexts (9, 11–16), but none of them in a really comprehensive manner. Here, we compared WES and WGS, in terms of detection rates and quality, for single-nucleotide variants (SNVs), small insertions/deletions (indels), and copy-number variants (CNVs) within the regions of the human genome covered by WES, using the most recent next-generation sequencing (NGS) technologies. We aimed to identify the most efficient and reliable approach for identifying these variants in coding regions of the genome, to define the optimal analytical filters for decreasing the frequency of false-positive variants, and to characterize the genes that were either hard to sequence by either approach or were poorly covered by WES kits.     

Monitoring gastric cancer progression with circulating tumour DNA

Background:

Circulating tumour DNA (ctDNA) is an emerging candidate biomarker for malignancies and may be useful for monitoring the disease status of gastric cancer.

Methods:

We performed targeted deep sequencing of plasma cell-free DNA (cfDNA) by massively parallel sequencing in patients with tumours harbouring TP53 mutations. The quantitative values of TP53-ctDNA during the clinical course were compared with the tumour status.

Results:

Three out of ten patients with TP53 mutations in primary tumours showed detectable TP53 mutation levels in preoperative cfDNA. Although the cfDNA concentrations were not always reflective of the disease course, the ctDNA fraction correlated with the disease status.

Conclusions:

ctDNA may serve as a useful biomarker to monitor gastric cancer progression and residual disease.     

Mutational analysis of pulmonary tumours with neuroendocrine features using targeted massive parallel sequencing: a comparison of a neglected tumour group

Background:

Lung cancer is the leading cause of cancer-related deaths worldwide. The typical and atypical carcinoid (TC and AC), the large-cell neuroendocrine carcinoma (LCNEC) and the small-cell lung cancers (SCLC) are subgroups of pulmonary tumours that show neuroendocrine differentiations. With the rising impact of molecular pathology in routine diagnostics the interest for reliable biomarkers, which can help to differentiate these subgroups and may enable a more personalised treatment of patients, grows.

Methods:

A collective of 70 formalin-fixed, paraffin-embedded (FFPE) pulmonary neuroendocrine tumours (17 TCs, 17 ACs, 19 LCNECs and 17 SCLCs) was used to identify biomarkers by high-throughput sequencing. Using the Illumina TruSeq Amplicon-Cancer Panel on the MiSeq instrument, the samples were screened for alterations in 221 mutation hot spots of 48 tumour-relevant genes.

Results:

After filtering >26 000 detected variants by applying strict algorithms, a total of 130 mutations were found in 29 genes and 49 patients. Mutations in JAK3, NRAS, RB1 and VHL1 were exclusively found in SCLCs, whereas the FGFR2 mutation was detected in LCNEC only. KIT, PTEN, HNF1A and SMO were altered in ACs. The SMAD4 mutation corresponded to the TC subtype. We prove that the frequency of mutations increased with the malignancy of tumour type. Interestingly, four out of five ATM-mutated patients showed an additional alteration in TP53, which was by far the most frequently altered gene (28 out of 130; 22%). We found correlations between tumour type and IASLC grade for ATM- (P=0.022; P=0.008) and TP53-mutated patients (P<0.001). Both mutated genes were also associated with lymph node invasion and distant metastasis (P⩽0.005). Furthermore, PIK3CA-mutated patients with high-grade tumours showed a reduced overall survival (P=0.040) and the mutation frequency of APC and ATM in high-grade neuroendocrine lung cancer patients was associated with progression-free survival (PFS) (P=0.020).

Conclusions:

The implementation of high-throughput sequencing for the analysis of the neuroendocrine lung tumours has revealed that, even if these tumours encompass several subtypes with varying clinical aggressiveness, they share a number of molecular features. An improved understanding of the biology of neuroendocrine tumours will offer the opportunity for novel approaches in clinical management, resulting in a better prognosis and prediction of therapeutic response.