Finding Driver Pathways in Cancer: Models and Algorithms

ABSTRACT: BACKGROUND: Cancer sequencing projects are now measuring somatic mutations in large numbers of cancer genomes. A key challenge in interpreting these data is to distinguish driver mutations, mutations important for cancer development, from passenger mutations that have accumulated in somatic cells but without functional consequences. A common approach to identify genes harboring driver mutations is a single gene test that identifies individual genes that are recurrently mutated in a significant number of cancer genomes. However, the power of this test is reduced by: (1) the necessity of estimating the background mutation rate (BMR) for each gene; (2) the mutational heterogeneity in most cancers meaning that groups of genes (e.g. pathways), rather than single genes, are the primary target of mutations. RESULTS: We investigate the problem of discovering driver pathways, groups of genes containing driver mutations, directly from cancer mutation data and without prior knowledge of pathways or other interactions between genes. We introduce two generative models of somatic mutations in cancer and study the algorithmic complexity of discovering driver pathways in both models. We show that a single gene test for driver genes is highly sensitive to the estimate of the BMR. In contrast, we show that an algorithmic approach that maximizes a straightforward measure of the mutational properties of a driver pathway successfully discovers these groups of genes without an estimate of the BMR. Moreover, this approach is also successful in the case when the observed frequencies of passenger and driver mutations are indistinguishable, a situation where single gene tests fail. CONCLUSIONS: Accurate estimation of the BMR is a challenging task. Thus, methods that do not require an estimate of the BMR, such as the ones we provide here, can give increased power for the discovery of driver genes.     

Clonal Architectures and Driver Mutations in Metastatic Melanomas

To reveal the clonal architecture of melanoma and associated driver mutations, whole genome sequencing (WGS) and targeted extension sequencing were used to characterize 124 melanoma cases. Significantly mutated gene analysis using 13 WGS cases and 15 additional paired extension cases identified known melanoma genes such as BRAF, NRAS, and CDKN2A, as well as a novel gene EPHA3, previously implicated in other cancer types. Extension studies using tumors from another 96 patients discovered a large number of truncation mutations in tumor suppressors (TP53 and RB1), protein phosphatases (e.g., PTEN, PTPRB, PTPRD, and PTPRT), as well as chromatin remodeling genes (e.g., ASXL3, MLL2, and ARID2). Deep sequencing of mutations revealed subclones in the majority of metastatic tumors from 13 WGS cases. Validated mutations from 12 out of 13 WGS patients exhibited a predominant UV signature characterized by a high frequency of C->T transitions occurring at the 3′ base of dipyrimidine sequences while one patient (MEL9) with a hypermutator phenotype lacked this signature. Strikingly, a subclonal mutation signature analysis revealed that the founding clone in MEL9 exhibited UV signature but the secondary clone did not, suggesting different mutational mechanisms for two clonal populations from the same tumor. Further analysis of four metastases from different geographic locations in 2 melanoma cases revealed phylogenetic relationships and highlighted the genetic alterations responsible for differential drug resistance among metastatic tumors. Our study suggests that clonal evaluation is crucial for understanding tumor etiology and drug resistance in melanoma.     

Search for the second Peutz-Jeghers syndrome locus: exclusion of the STK13, PRKCG,KLK10, and PSCD2 genes on chromosome 19 and the STK11IP gene on chromosome 2

Pathogenic mutations in the serine/threonine kinase STK11 (alias LKB1) cause Peutz-Jeghers syndrome (PJS) in most affected individuals. However, in a considerable number of PJS-patients mutations cannot be detected in STK11 suggesting genetic heterogeneity. One PJS family without STK11 mutations (PJS07) has previously been described with significant evidence for linkage to a second potential PJS locus on 19q13.3-->q13.4. In this study we investigated candidate genes within markers D19S180 and D19S254, since multipoint linkage analysis yielded significant LOD scores for this region in this family. Four genes in the region (cytohesin 2: PSCD2, kallikrein 10: KLK10, protein kinase C gamma: PRKCG, and serine/threonine kinase 13: STK13) potentially involved in growth inhibitory pathways or in the pathophysiology of can- cer, were considered as candidates. We first determined the genomic structure of the PSCD2 and PRKCG genes, and performed mutation analysis of all exons and exon-intron junctions of the four genes, in the PJS07 family. No pathogenic mutation was identified in these four genes in affected individuals. A very rare polymorphism resulting in a conserved amino acid change Lys to Arg was found in PSCD2. These data provide considerable evidence for exclusion of these four genes as candidates for the second locus on 19q13.3-->q13.4 in PJS. Finally, we also excluded the recently identified STK11-interacting protein gene (STK11IP, alias LIP1) mapped in 2q36 as candidate for PJS in the PJS07 family, although this could be a good candidate in other non-STK11/LKB1 families.     

Predicting mutant outcome by combining deep mutational scanning and machine learning

Deep mutational scanning provides unprecedented wealth of quantitative data regarding the functional outcome of mutations in proteins. A single experiment may measure properties (eg, structural stability) of numerous protein variants. Leveraging the experimental data to gain insights about unexplored regions of the mutational landscape is a major computational challenge. Such insights may facilitate further experimental work and accelerate the development of novel protein variants with beneficial therapeutic or industrially relevant properties. Here we present a novel, machine learning approach for the prediction of functional mutation outcome in the context of deep mutational screens. Using sequence (one-hot) features of variants with known properties, as well as structural features derived from models thereof, we train predictive statistical models to estimate the unknown properties of other variants. The utility of the new computational scheme is demonstrated using five sets of mutational scanning data, denoted “targets”: (a) protease specificity of APPI (amyloid precursor protein inhibitor) variants; (b-d) three stability related properties of IGBPG (immunoglobulin G-binding β1 domain of streptococcal protein G) variants; and (e) fluorescence of GFP (green fluorescent protein) variants. Performance is measured by the overall correlation of the predicted and observed properties, and enrichment—the ability to predict the most potent variants and presumably guide further experiments. Despite the diversity of the targets the statistical models can generalize variant examples thereof and predict the properties of test variants with both single and multiple mutations.     

The OSUMMER lines: A series of ultraviolet-accelerated NRAS-mutant mouse melanoma cell lines syngeneic to C57BL/6

An increasing number of cancer subtypes are treated with front-line immunotherapy. However, approaches to overcome primary and acquired resistance remain limited. Preclinical mouse models are often used to investigate resistance mechanisms, novel drug combinations, and delivery methods; yet most of these models lack the genetic diversity and mutational patterns observed in human tumors. Here we describe a series of 13 C57BL/6J melanoma cell lines to address this gap in the field. The Ohio State University-Moffitt Melanoma Exposed to Radiation (OSUMMER) cell lines are derived from mice expressing endogenous, melanocyte-specific, and clinically relevant Nras driver mutations (Q61R, Q61K, or Q61L). Exposure of these animals to a single, non-burning dose of ultraviolet B accelerates the onset of spontaneous melanomas with mutational patterns akin to human disease. Furthermore, in vivo irradiation selects against potent tumor antigens, which could prevent the outgrowth of syngeneic cell transfers. Each OSUMMER cell line possesses distinct in vitro growth properties, trametinib sensitivity, mutational signatures, and predicted antigenicity. Analysis of OSUMMER allografts shows a correlation between strong, predicted antigenicity and poor tumor outgrowth. These data suggest that the OSUMMER lines will be a valuable tool for modeling the heterogeneous responses of human melanomas to targeted and immune-based therapies.     

VarWalker: Personalized Mutation Network Analysis of Putative Cancer Genes from Next-Generation Sequencing Data

A major challenge in interpreting the large volume of mutation data identified by next-generation sequencing (NGS) is to distinguish driver mutations from neutral passenger mutations to facilitate the identification of targetable genes and new drugs. Current approaches are primarily based on mutation frequencies of single-genes, which lack the power to detect infrequently mutated driver genes and ignore functional interconnection and regulation among cancer genes. We propose a novel mutation network method, VarWalker, to prioritize driver genes in large scale cancer mutation data. VarWalker fits generalized additive models for each sample based on sample-specific mutation profiles and builds on the joint frequency of both mutation genes and their close interactors. These interactors are selected and optimized using the Random Walk with Restart algorithm in a protein-protein interaction network. We applied the method in >300 tumor genomes in two large-scale NGS benchmark datasets: 183 lung adenocarcinoma samples and 121 melanoma samples. In each cancer, we derived a consensus mutation subnetwork containing significantly enriched consensus cancer genes and cancer-related functional pathways. These cancer-specific mutation networks were then validated using independent datasets for each cancer. Importantly, VarWalker prioritizes well-known, infrequently mutated genes, which are shown to interact with highly recurrently mutated genes yet have been ignored by conventional single-gene-based approaches. Utilizing VarWalker, we demonstrated that network-assisted approaches can be effectively adapted to facilitate the detection of cancer driver genes in NGS data.     

Low immunogenicity of common cancer hot spot mutations resulting in false immunogenic selection signals

Cancer is driven by somatic mutations that result in a cellular fitness advantage. This selective advantage is expected to be counterbalanced by the immune system when these driver mutations simultaneously lead to the generation of neoantigens, novel peptides that are presented at the cancer cell membrane via HLA molecules from the MHC complex. The presentability of these peptides is determined by a patient’s MHC genotype and it has been suggested that this results in MHC genotype-specific restrictions of the oncogenic mutational landscape. Here, we generated a set of virtual patients, each with an identical and prototypical MHC genotype, and show that the earlier reported HLA affinity differences between observed and unobserved mutations are unrelated to MHC genotype variation. We demonstrate how these differences are secondary to high frequencies of 13 hot spot driver mutations in 6 different genes. Several oncogenic mechanisms were identified that lower the peptides’ HLA affinity, including phospho-mimicking substitutions in BRAF, destabilizing tyrosine mutations in TP53 and glycine-rich mutational contexts in the GTP-binding KRAS domain. In line with our earlier findings, our results emphasize that HLA affinity predictions are easily misinterpreted when studying immunogenic selection processes.     

Guanine Holes Are Prominent Targets for Mutation in Cancer and Inherited Disease

Single base substitutions constitute the most frequent type of human gene mutation and are a leading cause of cancer and inherited disease. These alterations occur non-randomly in DNA, being strongly influenced by the local nucleotide sequence context. However, the molecular mechanisms underlying such sequence context-dependent mutagenesis are not fully understood. Using bioinformatics, computational and molecular modeling analyses, we have determined the frequencies of mutation at G•C bp in the context of all 64 5′-NGNN-3′ motifs that contain the mutation at the second position. Twenty-four datasets were employed, comprising >530,000 somatic single base substitutions from 21 cancer genomes, >77,000 germline single-base substitutions causing or associated with human inherited disease and 16.7 million benign germline single-nucleotide variants. In several cancer types, the number of mutated motifs correlated both with the free energies of base stacking and the energies required for abstracting an electron from the target guanines (ionization potentials). Similar correlations were also evident for the pathological missense and nonsense germline mutations, but only when the target guanines were located on the non-transcribed DNA strand. Likewise, pathogenic splicing mutations predominantly affected positions in which a purine was located on the non-transcribed DNA strand. Novel candidate driver mutations and tissue-specific mutational patterns were also identified in the cancer datasets. We conclude that electron transfer reactions within the DNA molecule contribute to sequence context-dependent mutagenesis, involving both somatic driver and passenger mutations in cancer, as well as germline alterations causing or associated with inherited disease.     

Cancer driver mutations: predictions and reality

Cancer cells accumulate many genetic alterations throughout their lifetime, but only a few of them drive cancer progression, termed driver mutations. Driver mutations may vary between cancer types and patients, can remain latent for a long time and become drivers at particular cancer stages, or may drive oncogenesis only in conjunction with other mutations. The high mutational, biochemical, and histological tumor heterogeneity makes driver mutation identification very challenging. In this review we summarize recent efforts to identify driver mutations in cancer and annotate their effects. We underline the success of computational methods to predict driver mutations in finding novel cancer biomarkers, including in circulating tumor DNA (ctDNA). We also report on the boundaries of their applicability in clinical research.     

Nine novel germline mutations of STK11 in ten families with Peutz-Jeghers syndrome

Peutz-Jeghers Syndrome (PJS) is an autosomal dominant hereditary disease characterized by hamartomatous polyposis involving the entire bowel. Recently STK11, a gene bearing a mutation responsible for PJS, was isolated. We investigated the entire coding region of STK11 in 15 unrelated PJS families by the PCR-SSCP (polymerase chain reaction-single strand conformation polymorphism) method and PCR-direct sequence analysis, and found nine different, novel mutations among ten of those families. One nonsense mutation and five different frameshift mutations (two families carried the same mutation), all of which would cause truncation of the gene product, were found in seven families; mutations found in five families were clustered within exon 6. Among these five mutations, three occurred at the mononucleotide-repeat region (CCCCCC) of codons 279–281, suggesting that this region is likely to be a mutational hotspot of this gene. One of the remaining three families carried a 3-bp in-frame deletion that would eliminate an asparagine residue within a kinase domain of the product; the other two carried intronic mutations at or adjacent to the consensus dinucleotide sequences of splice-acceptor or -donor sites, which were likely to lead to aberrant splicing. Received: 9 March 1998 / Accepted: 1 May 1998     

Comparing mutational pathways to lopinavir resistance in HIV-1 subtypes B versus C

Although combination antiretroviral therapies seem to be effective at controlling HIV-1 infections regardless of the viral subtype, there is increasing evidence for subtype-specific drug resistance mutations. The order and rates at which resistance mutations accumulate in different subtypes also remain poorly understood. Most of this knowledge is derived from studies of subtype B genotypes, despite not being the most abundant subtype worldwide. Here, we present a methodology for the comparison of mutational networks in different HIV-1 subtypes, based on Hidden Conjunctive Bayesian Networks (H-CBN), a probabilistic model for inferring mutational networks from cross-sectional genotype data. We introduce a Monte Carlo sampling scheme for learning H-CBN models for a larger number of resistance mutations and develop a statistical test to assess differences in the inferred mutational networks between two groups. We apply this method to infer the temporal progression of mutations conferring resistance to the protease inhibitor lopinavir in a large cross-sectional cohort of HIV-1 subtype C genotypes from South Africa, as well as to a data set of subtype B genotypes obtained from the Stanford HIV Drug Resistance Database and the Swiss HIV Cohort Study. We find strong support for different initial mutational events in the protease, namely at residue 46 in subtype B and at residue 82 in subtype C. The inferred mutational networks for subtype B versus C are significantly different sharing only five constraints on the order of accumulating mutations with mutation at residue 54 as the parental event. The results also suggest that mutations can accumulate along various alternative paths within subtypes, as opposed to a unique total temporal ordering. Beyond HIV drug resistance, the statistical methodology is applicable more generally for the comparison of inferred mutational networks between any two groups.     

An Evolutionary Approach for Identifying Driver Mutations in Colorectal Cancer

The traditional view of cancer as a genetic disease that can successfully be treated with drugs targeting mutant onco-proteins has motivated whole-genome sequencing efforts in many human cancer types. However, only a subset of mutations found within the genomic landscape of cancer is likely to provide a fitness advantage to the cell. Distinguishing such “driver” mutations from innocuous “passenger” events is critical for prioritizing the validation of candidate mutations in disease-relevant models. We design a novel statistical index, called the Hitchhiking Index, which reflects the probability that any observed candidate gene is a passenger alteration, given the frequency of alterations in a cross-sectional cancer sample set, and apply it to a mutational data set in colorectal cancer. Our methodology is based upon a population dynamics model of mutation accumulation and selection in colorectal tissue prior to cancer initiation as well as during tumorigenesis. This methodology can be used to aid in the prioritization of candidate mutations for functional validation and contributes to the process of drug discovery.     

Involvement of 5-methylcytosine in sunlight-induced mutagenesis.

In human skin cancers, more than 30 % of all mutations in the p53 gene are transitions at dipyrimidines within the sequence context CpG, i.e. 5'-TCG and 5'-CCG, found at several mutational hotspots. Since CpGs are methylated along the p53 gene, these mutations may be derived from solar UV-induced pyrimidine dimers forming at sequences that contain 5-methylcytosine. In Xorder to define the contribution of 5-methylcytosine to sunlight-induced mutations, we have used mouse fibroblasts containing the CpG-methylated lacI transgene as a mutational target. We sequenced 182 UVC (254 nm UV)-induced mutations and 170 mutations induced by a solar UV simulator, along with 75 mutations in untreated cells. Only a few of the mutations in untreated cells were transitions at dipyrimidines, but more than 95% of the UVC and solar irradiation-induced mutations were targeted to dipyrimidine sites, the majority being transitions. After UVC irradiation, 6% of the base substitutions were at dipyrimidines containing 5-methylcytosine and only 2.2% of all mutations were transitions within this sequence context. However, 24% of the solar light-induced mutations were at dipyrimidines that contain 5-methylcytosine and most of them were transitions. Two sunlight-induced mutational hotspots at methylated CpGs correlated with sequences that form the highest levels of cyclobutane pyrimidine dimers after irradiation with sunlight but not with UVC. The data indicate that dipyrimidines that contain 5-methylcytosine are preferential targets for sunlight-induced mutagenesis in cultured mammalian cells, thus explaining the large proportion of p53 mutations at such sites in skin tumors in vivo.     

The genetic heterogeneity and mutational burden of engineered melanomas in zebrafish models

Background

Melanoma is the most deadly form of skin cancer. Expression of oncogenic BRAF or NRAS, which are frequently mutated in human melanomas, promote the formation of nevi but are not sufficient for tumorigenesis. Even with germline mutated p53, these engineered melanomas present with variable onset and pathology, implicating additional somatic mutations in a multi-hit tumorigenic process.

Results

To decipher the genetics of these melanomas, we sequence the protein coding exons of 53 primary melanomas generated from several BRAF^V600E or NRAS^Q61K driven transgenic zebrafish lines. We find that engineered zebrafish melanomas show an overall low mutation burden, which has a strong, inverse association with the number of initiating germline drivers. Although tumors reveal distinct mutation spectrums, they show mostly C > T transitions without UV light exposure, and enrichment of mutations in melanogenesis, p53 and MAPK signaling. Importantly, a recurrent amplification occurring with pre-configured drivers BRAF^V600E and p53^-/- suggests a novel path of BRAF cooperativity through the protein kinase A pathway.

Conclusion

This is the first analysis of a melanoma mutational landscape in the absence of UV light, where tumors manifest with remarkably low mutation burden and high heterogeneity. Genotype specific amplification of protein kinase A in cooperation with BRAF and p53 mutation suggests the involvement of melanogenesis in these tumors. This work is important for defining the spectrum of events in BRAF or NRAS driven melanoma in the absence of UV light, and for informed exploitation of models such as transgenic zebrafish to better understand mechanisms leading to human melanoma formation.     

Molecular Typing of Lung Adenocarcinoma on Cytological Samples Using a Multigene Next Generation Sequencing Panel

Identification of driver mutations in lung adenocarcinoma has led to development of targeted agents that are already approved for clinical use or are in clinical trials. Therefore, the number of biomarkers that will be needed to assess is expected to rapidly increase. This calls for the implementation of methods probing the mutational status of multiple genes for inoperable cases, for which limited cytological or bioptic material is available. Cytology specimens from 38 lung adenocarcinomas were subjected to the simultaneous assessment of 504 mutational hotspots of 22 lung cancer-associated genes using 10 nanograms of DNA and Ion Torrent PGM next-generation sequencing. Thirty-six cases were successfully sequenced (95%). In 24/36 cases (67%) at least one mutated gene was observed, including EGFR, KRAS, PIK3CA, BRAF, TP53, PTEN, MET, SMAD4, FGFR3, STK11, MAP2K1. EGFR and KRAS mutations, respectively found in 6/36 (16%) and 10/36 (28%) cases, were mutually exclusive. Nine samples (25%) showed concurrent alterations in different genes. The next-generation sequencing test used is superior to current standard methodologies, as it interrogates multiple genes and requires limited amounts of DNA. Its applicability to routine cytology samples might allow a significant increase in the fraction of lung cancer patients eligible for personalized therapy.     

The YUMM lines: a series of congenic mouse melanoma cell lines with defined genetic alterations

The remarkable success of immune therapies emphasizes the need for immune‐competent cancer models. Elegant genetically engineered mouse models of a variety of cancers have been established, but their effective use is limited by cost and difficulties in rapidly generating experimental data. Some mouse cancer cell lines are transplantable to immunocompetent host mice and have been utilized extensively to study cancer immunology. Here, we describe the Yale University Mouse Melanoma (YUMM) lines, a comprehensive system of mouse melanoma cell lines that are syngeneic to C57BL/6, have well‐defined human‐relevant driver mutations, and are genomically stable. This will be a useful tool for the study of tumor immunology and genotype‐specific cancer biology.     

Influence of uvrB and pKM101 on the spectrum of spontaneous, UV- and gamma-ray-induced base substitutions that revert hisG46 in Salmonella typhimurium

Oligonucleotide probes were used to identify base substitutions in 1089 revertants of hisG46 in Salmonella typhimurium that arose spontaneously or following irradiation with UV- or gamma-rays. The hisG46 allele, carrying a mutant CCC codon (Pro) in place of the wild-type codon CTC (Leu69) reverted via 6 distinguishable mutational events--C to T transitions at codon sites 1 or 2, C to A or C to G transversions at codon site 1, C to A at codon site 2, and an extragenic suppressor mutation. The distribution of hisG46 revertants differed among treatments and was influenced by the DNA-repair capacity of the bacteria. Plasmid pKM101 enhanced the frequencies of both spontaneous and induced mutations; transversion events were enhanced more efficiently by pKM101 than were transition events. Compared to Uvr+ bacteria, Uvr- bacteria had higher frequencies of spontaneous and induced mutations; transition mutations were enhanced more efficiently than were transversion mutations. The influence of DNA-repair activities on the mutational spectra provides some insights on the origins of spontaneous and UV-induced mutations.     

Analysis of Dysgenesis-Induced Lethal Mutations on the X Chromosome of a Q Strain of DROSOPHILA MELANOGASTER

The Q strain known as v6 was tested for its ability to induce X-linked lethal mutations in male and female hybrids from crosses with M strains in the P-M system of hybrid dysgenesis. All measurements of the mutation rate were made on the X chromosome derived from the v6 strain. The lethal rate for young hybrid males from the cross M female X v6 male was 1.11% per chromosome. For older males, it was only 0.44%, suggesting that there is less mutational or more repair activity in the germ cells of the older males or that mutant cells are selectively eliminated as the hybrid males age. The lethal rate for hybrid females from comparable crosses was approximately the same for both ages that were tested. However, it was substantially less than the rate for the hybrid males--only 0.26% per chromosome. Genetically identical hybrid females from reciprocal crosses also showed a low mutation rate, 0.13% per chromosome. Again, there was no difference between young and old flies. Mapping experiments established that most of the lethal mutations that were recovered from the male and female hybrids were located in two regions on the X chromosome, one between bands 14B13 and 15A9 , the other between bands 19A1 and 20A , which encompasses the maroonlike locus. More refined mapping of the lethals in the maroonlike region demonstrated that the vast majority of these affected a single gene located in band 19C4 . Cytological analysis of the lethal chromosomes revealed that several carried rearrangements, including inversions, duplications and deficiencies. Chromosome breakage occurred primarily in bands 14D1 -3 and 18F- 20A , and most of the breaks in the latter segment were located in 19C . However, rearrangements involving 19C and mutations of the gene in 19C4 were mutually exclusive events. In situ hybridization of a P element probe to the chromosomes of v6 demonstrated that P elements reside at a minimum of five sites on the X chromosome. These P element sites correspond to the mutational and breakage hot spots on that chromosome. The combined genetic and cytological data imply that most of the X-linked lethal mutations that occur in M X v6 hybrids are due to local P element action. Consideration of these and other data suggest that v6 is a weak P strain in the P-M system of hybrid dysgenesis and that other Q strains might also be regarded in this way.     

Evaluating the within-host fitness effects of mutations fixed during virus adaptation to different ecotypes of a new host

The existence of genetic variation for resistance in host populations is assumed to be essential to the spread of an emerging virus. Models predict that the rate of spread slows down with the increasing frequency and higher diversity of resistance alleles in the host population. We have been using the experimental pathosystem Arabidopsis thaliana—tobacco etch potyvirus (TEV) to explore the interplay between genetic variation in host''s susceptibility and virus diversity. We have recently shown that TEV populations evolving in A. thaliana ecotypes that differ in susceptibility to infection gained within-host fitness, virulence and infectivity in a manner compatible with a gene-for-gene model of host–parasite interactions: hard-to-infect ecotypes were infected by generalist viruses, whereas easy-to-infect ecotypes were infected by every virus. We characterized the genomes of the evolved viruses and found cases of host-driven convergent mutations. To gain further insights in the mechanistic basis of this gene-for-gene model, we have generated all viral mutations individually as well as in specific combinations and tested their within-host fitness effects across ecotypes. Most of these mutations were deleterious or neutral in their local ecotype and only a very reduced number had a host-specific beneficial effect. We conclude that most of the mutations fixed during the evolution experiment were so by drift or by selective sweeps along with the selected driver mutation. In addition, we evaluated the ruggedness of the underlying adaptive fitness landscape and found that mutational effects were mostly multiplicative, with few cases of significant epistasis.     

Evolutionary accessibility of mutational pathways

Functional effects of different mutations are known to combine to the total effect in highly nontrivial ways. For the trait under evolutionary selection ('fitness'), measured values over all possible combinations of a set of mutations yield a fitness landscape that determines which mutational states can be reached from a given initial genotype. Understanding the accessibility properties of fitness landscapes is conceptually important in answering questions about the predictability and repeatability of evolutionary adaptation. Here we theoretically investigate accessibility of the globally optimal state on a wide variety of model landscapes, including landscapes with tunable ruggedness as well as neutral 'holey' landscapes. We define a mutational pathway to be accessible if it contains the minimal number of mutations required to reach the target genotype, and if fitness increases in each mutational step. Under this definition accessibility is high, in the sense that at least one accessible pathway exists with a substantial probability that approaches unity as the dimensionality of the fitness landscape (set by the number of mutational loci) becomes large. At the same time the number of alternative accessible pathways grows without bounds. We test the model predictions against an empirical 8-locus fitness landscape obtained for the filamentous fungus Aspergillus niger. By analyzing subgraphs of the full landscape containing different subsets of mutations, we are able to probe the mutational distance scale in the empirical data. The predicted effect of high accessibility is supported by the empirical data and is very robust, which we argue reflects the generic topology of sequence spaces. Together with the restrictive assumptions that lie in our definition of accessibility, this implies that the globally optimal configuration should be accessible to genome wide evolution, but the repeatability of evolutionary trajectories is limited owing to the presence of a large number of alternative mutational pathways.