SEARCH · bioRxiv Science

Results for “genetics”

Search indexed bioRxiv preprints in genomics, neuroscience, cell biology and bioinformatics. Read source abstracts and check manuscript versions; preprints are not peer reviewed.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,639 records · Page 91Linked to original sources

bioRxiv2016

A computational method for the investigation of multistable systems and its application to genetic switches

Genetic switches exhibit multistability, form the basis of epigenetic memory, and are found in natural decision making systems, such as cell fate determination in developmental pathways. Synthetic genetic switches can be used for recording the presence of different environmental signals, for changing phenotype using synthetic inputs and as building blocks for higher-level sequential logic circuits. Understanding how multistable switches can be constructed and how they function within larger biological systems is therefore key to synthetic biology. Here we present a new computational tool, called StabilityFinder, that takes advantage of sequential Monte Carlo methods to identify regions of parameter space capable of producing multistable behaviour, while handling uncertainty in biochemical rate constants and initial conditions. The algorithm works by clustering trajectories in phase space, and iteratively minimizing a distance metric. Here we examine a collection of models of genetic switches, ranging from the deterministic Gardner toggle switch to stochastic models containing different positive feedback connections. We uncover the design principles behind making bistable, tristable and quadristable switches, and find that rate of gene expression is a key parameter. We demonstrate the ability of the framework to examine more complex systems and examine the design principles of a three gene switch. Our framework allows us to relax the assumptions that are often used in genetic switch models and we show that more complex abstractions are still capable of multistable behaviour. Our results suggest many ways in which genetic switches can be enhanced and offer designs for the construction of novel switches. Our analysis also highlights subtle changes in correlation of experimentally tunable parameters that can lead to bifurcations in deterministic and stochastic systems. Overall we demonstrate that StabilityFinder will be a valuable tool in the future design and construction of novel gene networks.

synthetic biology↗

bioRxiv2016

Genetic footprint of population fragmentation and contemporary collapse in a freshwater cetacean

Understanding demographic trends and patterns of gene flow in an endangered species is crucial for devising conservation strategies. Here, we examined the extent of population structure and recent evolution of the critically endangered Yangtze finless porpoise (Neophocaena asiaeorientalis asiaeorientalis). By analysing genetic variation at the mitochondrial and nuclear microsatellite loci for 148 individuals, we identified three populations along the Yangtze River, each one connected to a group of admixed ancestry. Each population displayed extremely low genetic diversity, consistent with extremely small effective size ([≤]92 individuals). Habitat degradation and distribution gaps correlated with highly asymmetric gene-flow that was inefficient in maintaining connectivity between populations. Genetic inferences of historical demography revealed that the populations in the Yangtze descended from a small number of founders colonizing the river from the sea during the last Ice Age. The colonization was followed by a rapid population split during the last millennium predating the Chinese Modern Economy Development. However, genetic diversity showed a clear footprint of population contraction over the last 50 years leaving only ~2% of the pre-collapsed size, consistent with the population collapses reported from field studies. This genetic perspective provides background information for devising mitigation strategies to prevent this species from extinction.

evolutionary biology↗

bioRxiv2016

Sex matters in Massive Parallel Sequencing: Evidence for biases in genetic parameter estimation and investigation of sex determination systems

Using massively parallel sequencing data from two species with different life history traits -- American lobster (Homarus americanus) and Arctic Char (Salvelinus alpinus) -- we highlighted how an unbalanced sex ratio in the samples combined with a few sex-linked markers may lead to false interpretations of population structure and thus to potentially erroneous management recommendations. Multivariate analyses revealed two genetic clusters that separated males and females instead of showing the expected pattern of genetic differentiation among ecologically divergent (inshore vs. offshore in lobster) or geographically distant (east vs. west in Arctic Char) sampling locations. We created several subsamples artificially varying the sex ratio in the inshore/offshore and east/west groups, and then demonstrated that significant genetic differentiation could be observed despite panmixia for lobster, and that Fst values were overestimated for Arctic Char. This pattern was due to 12 and 94 sex-linked markers driving differentiation for lobster and Arctic Char, respectively. Removing sex-linked markers led to nonsignificant genetic structure (lobster) and a more accurate estimation of Fst (Arctic Char). We further characterized the putative functions of sex-linked markers. Given that only 9.6% of all marine/diadromous population genomic studies to date reported sex information, we urge researchers to collect and consider individual sex information. In summary, we argue that sex information is useful to (i) control sex ratio in sampling, (ii) overcome \"sex-ratio bias\" that can lead to spurious genetic differentiation signals and (iii) fill knowledge gaps regarding sex determining systems.

genomics↗

bioRxiv2017

The systematic conservation planning for intraspecific genetic diversity.

Intraspecific diversity informs the demographic and evolutionary histories of populations, and should be a main conservation target. Although approaches exist for identifying relevant biological conservation units, attempts to identify priority conservation areas for intraspecific diversity are scarce, especially within a multi-specific framework. We used neutral molecular data on six European freshwater fish species (Squalius cephalus, Phoxinus phoxinus, Barbatula barbatula, Gobio occitaniae, Leuciscus burdigalensis and Parachondrostoma toxostoma) sampled at the riverscape scale (i.e. the Garonne-Dordogne River basin, France) to determine hot- and cold-spots of genetic diversity, and to identify priority conservation areas using a systematic conservation planning approach. We demonstrate that systematic conservation planning is efficient for identifying priority areas representing a predefined part of the total genetic diversity of a whole landscape. With the exception of private allelic richness, classical genetic diversity indices (allelic richness, genetic uniqueness) were poor predictors for identifying priority areas. Moreover, we identified weak surrogacies among conservation solutions found for each species, implying that conservation solutions are highly species-specific. Nonetheless, we showed that priority areas identified using intraspecific genetic data from multiple species provide more effective conservation solutions than areas identified for single species or on the basis of traditional taxonomic criteria.

evolutionary biology↗

bioRxiv2017

Heritability of the structures and 13C fractionation in tomato leaf wax alkanes: a genetic model system to inform paleoenvironmental reconstructions

Leaf wax n-alkanes are broadly used to reconstruct paleoenvironmental information. However, the utility of the n-alkane paleoclimate proxy is modulated by the extent to which genetic as well as environmental factors influence the structural and isotopic variability of leaf waxes. In paleoclimate applications, there is an implicit assumption that most variation of leaf wax traits through a time series can be attributed to environmental change and that biological sources of variability within plant communities are small. For example, changes in hydrology affect the {delta}2 H of waxes though rainwater and the {delta}13C of leaf waxes by changing plant communities (i.e., C3 versus C4 input). Here we test the assumption of little genetic control over 5 C variation of leaf wax by presenting the results of an experimental greenhouse growth study in which we estimate the role of genetic variability on structural and isotopic leaf wax traits in a set of 76 introgression lines (ILs) between two interfertile Solanum (tomato) species: S. lycopersicum cv M82 (hereafter cv M82) and S. pennellii. We found that the leaves of S. pennellii, a wild desert tomato relative, produces significantly more iso-alkanes than cv M82, a domesticated tomato cultivar adapted to water-replete conditions; we introduce a methylation index to summarize the ratio of branched (iso- and anteiso-) to total alkanes. Between S. pennellii and cv M82, the iso-alkanes were found to be enriched in 13C by 1.2-1.4%o over n-alkanes. By modeling our results from the ILs, we report the broad-sense heritability values (H2) of leaf wax traits to describe the degree to which genetic variation contributes to variation of these traits. Individual carbon isotope values of alkanes are of low heritability (H2 = 0.13-0.19), suggesting that {delta}13C of leaf waxes from this study are strongly influenced by environmental variance, which supports the interpretation that variation in the 5 C of wax compounds recorded in sediments reflects paleohydrological changes. Average chain length (ACL) values of n-alkanes are of intermediate heritability (H2 = 0.30), suggesting that ACL values are strongly influenced by genetic cues.

plant biology↗

bioRxiv2017

Maintaining their genetic distance; limited gene flow between widely hybridising species of Geum with contrasting mating systems

Mating system transition from outcrossing to selfing frequently gives rise to sister lineages with contrasting outcrossing rates. The evolutionary fate of such lineages depends on the extent to which they exchange genes. We measured gene flow between outcrossing Geum rivale and selfing G. urbanum, two sister species derived by mating system transition, which frequently hybridise. A draft genome was generated for G. urbanum and used to develop dd-RAD data scorable in both species. Coalescent analysis of RAD data from allopatric populations indicated that the two species diverged 2-3 Mya, and that long term gene flow between them has been very low (M=0.04). G. rivale showed greater genetic diversity in sympatry than allopatry, but genetic divergence between species was no lower in sympatry than allopatry, providing little evidence for recent introgression. Clustering of genotypes revealed that, apart from four early generation hybrids, individuals in sympatric populations fell into two genetically distinct groups with <1% admixture that corresponded exactly to their morphological species classification. Although our data suggest limited gene flow, we observed joint segregation of two putatively introgressed SNPs in G. urbanum populations that was associated with significant morphological variation; this provides tentative evidence for rare introduction of novel genetic diversity by interspecific gene flow. Our results indicate that despite frequent hybridisation, genetic exchange between G. rivale and G. urbanum has been very limited throughout their evolutionary history.

evolutionary biology↗

bioRxiv2017

Decoding directional genetic dependencies through orthogonal CRISPR/Cas screens

Genetic interaction studies are a powerful approach to identify functional interactions between genes. This approach can reveal networks of regulatory hubs and connect uncharacterised genes to well-studied pathways. However, this approach has previously been limited to simple gene inactivation studies. Here, we present an orthogonal CRISPR/Cas-mediated genetic interaction approach that allows the systematic activation of one gene while simultaneously knocking out a second gene in the same cell. We have developed this concept into a quantitative and scalable combinatorial screening platform that allows the parallel interrogation of hundreds of thousands of genetic interactions. We demonstrate that the established platform works robustly to uncover genetic interactions in human cancer cells and to interpret the direction of the flow of genetic information.

molecular biology↗

bioRxiv2017

Elucidating dispersal ecology of reclusive species through genetic analyses of parentage and relatedness: the island night lizard (Xantusia riversiana) as a case study.

Characterizing dispersal and movement patterns are vital to understanding the evolutionary ecology of species. For many reclusive species, such as reptiles, the observation of direct dispersal may be difficult or intractable. However, dispersal distances and patterns may be characterized through indirect genetic methods. We used genetic and capture data from the island night lizard (Xantusia riversiana) to estimate natal dispersal distances through indirect genetic methods, characterize movement and space use patterns, and compare these distances to previous estimates made from more traditional ecological approaches. We found that indirect estimates of natal dispersal were greater than previous field-based estimates of individual displacement of 3-5 m. Parent-offspring differences had a mean of approximately 14 m on Santa Barbara Island (SBI) and 41 m on San Clemente Island (SCI) whereas Wrights {sigma} was estimated at 16 m on SBI and 20 m for SCI. Spatial autocorrelation with correlograms of Morans I revealed large differences in the scale of autocorrelation between islands (SBI=375 m, SCI=1,813 m). Interpretation of these distances as average per generation distance of gene flow was incongruent with parentage analyses and {sigma}. We also used variograms to evaluate the range of spatial autocorrelation among two inter-individual genetic differences. The range of spatial autocorrelation again identified different scales on the two islands (102 - 169 m on SBI and 955 - 1,424 m on SCI). No evidence of sex-biased dispersal was found on either island. However, a permutation logistic regression revealed that related individuals >0.8 years old were more likely to be captured together on both islands. Overall, our findings suggest that field-based estimates of individual displacement within this species may underestimate genetic dispersal. We suggest indirect inferences of natal dispersal distances should focus on parentage analyses and Wrights {sigma} for parameter estimation of individual movement, whereas the ranges identified by spatial autocorrelation and variograms are likely to be relevant at the metapopulation or patch scales. Furthermore, characterization of capture patterns and relatedness revealed kin-affiliative behavior in X. riversiana, which may be indicative of delayed dispersal and cryptic sociality. These results highlight the power of parentage- and relatedness-based analyses for characterizing aspects of the movement ecology of reclusive species that may be difficult to observe directly. These data can then be leveraged to support future conservation and population modeling efforts and assess extinction risks and management strategies.

ecology↗

bioRxiv2017

Refactoring the Genetic Code for Increased Evolvability

The standard genetic code is robust to mutations and base-pairing errors during transcription and translation. Point mutations are most likely to be synonymous or preserve the chemical properties of the original amino acid. Saturation mutagenesis experiments suggest that in some cases the best performing mutant requires a replacement of more than a single nucleotide within a codon. These replacements are essentially inaccessible to common error-based laboratory engineering techniques that alter single nucleotide per mutation event, due to the extreme rarity of adjacent mutations. In this theoretical study, we suggest a radical reordering of the genetic code that maximizes the mutagenic potential of single nucleotide replacements. We explore several possible genetic codes that allow a greater degree of accessibility to the mutational landscape and may result in a hyper-evolvable organism serving as an ideal platform for directed evolution experiments. We then conclude by evaluating potential applications for recoded organisms within the synthetic biology field.\n\nSignificance StatementThe conservative nature of the genetic code prevents bioengineers from efficiently accessing the full mutational landscape of a gene using common error-prone methods. Here we present two computational approaches to generate alternative genetic codes with increased accessibility. These new codes allow mutational transition to a larger pool of amino acids and with a greater degree of chemical differences, using a single nucleotide replacement within the codon, thus increasing evolvability both at the single gene and at the genome levels. Given the widespread use of these techniques for strain and protein improvement along with more fundamental evolutionary biology questions, the use of recoded organisms that maximize evolvability should significantly improve the efficiency of directed evolution, library generation and fitness maximization.

bioengineering↗

bioRxiv2017

Contribution Of Genetic Variation And Developmental Stage To Methylome Dynamics In Myeloid Differentiation

DNA methylation is important to establish a cells developmental identity. It also modulates cellular responses to endogenous developmental stimuli or environmental changes. We designed an in vitro myeloid differentiation model to analyze the genetic and developmental contribution to methylome dynamics using whole-genome bisulfide sequencing and transcriptome sequencing. Using a recursive partitioning approach, we identified 34,502 differentially methylated regions (DMRs) associated with genetic background and/or developmental stimuli. Specifically, 23,792 DMRs (69%) were significantly associated with inter-individual variations, of which 82% were associated with genetic polymorphisms in cis. Notably, inter-individual variations further modified 57 of 212 (26%) developmental DMRs with transcriptomic responses. Our study presents a novel analytical approach to determine the bona fide genetic contribution embedded in outlier patterns of CpG-SNPs in individual methylomes. This approach can be used to study genetic and epigenetic mechanisms underlying differential responses to developmental stimuli, environmental changes, and inter-individual differences in drug responses.

genomics↗

bioRxiv2017

Sex Differentiation In Grayling (Salmonidae) Goes Through An All-Male Stage And Is Delayed In Genetic Males Who Instead Grow Faster

Fish can be threatened by distorted sex ratios that arise during sex differentiation. It is therefore important to understand sex determination and differentiation, especially in river-dwelling fish that are often exposed to environmental factors that may interfere with sex differentiation. However, sex differentiation is not sufficiently understood in keystone taxa such as the Thymallinae, one of the three salmonid subfamilies. Here we study a wild grayling (Thymallus thymallus) population that suffers from distorted sex ratios. We found sex determination in the wild and in captivity to be genetic and linked to the sdY locus. We therefore studied sex-specific gene expression in embryos and early larvae that were bred and raised under different experimental conditions, and we studied gonadal morphology in five monthly samples taken after hatching. Significant sex-specific changes in gene expression (affecting about 25,000 genes) started around hatching. Gonads were still undifferentiated three weeks after hatching, but about half of the fish showed immature testes around seven weeks after hatching. Over the next few months, this phenotype was mostly replaced by the \"testis-to-ovary\" or \"ovaries\" phenotypes. The gonads of the remaining fish, i.e. approximately half of the fish in each sampling period, remained undifferentiated until six months after fertilization. Genetic sexing of the last two samples revealed that fish with undifferentiated gonads were all males, who, by that time, were on average larger than the genetic females (verified in 8-months old juveniles raised in another experiment). Only 12% of the genetic males showed testicular tissue six months after fertilization. We conclude that sex differentiation starts around hatching, goes through an all-male stage for both sexes (which represents a rare case of \"undifferentiated\" gonochoristic species that usually go through an all-female stage), and is delayed in males who, instead of developing their gonads, grow faster than females during these juvenile stages.\n\nAuthor contributionMRR and CW initiated the project. DM, OS, AU, LMC, LW, and CW sampled the adult fish, did the experimental in vitro fertilizations, and prepared the embryos for experimental rearing in the laboratory. All further manipulations on the embryos and the larvae were done by DM, OS, AU, LMC, and LW. The RNA-seq data were analyzed by OS, JR, and MRR, the histological analyses were done by DM, supervised by SK, and the molecular genetic sexing was performed by DM, OS, AU, and KBM. DM, OS, and CW performed the remaining statistical analyses and wrote the first version of the manuscript that was then critically revised by all other authors.

developmental biology↗

bioRxiv2017

Interdependence, Reflexivity, Fidelity, Impedance Matching, And The Evolution Of Genetic Coding

Genetic coding is generally thought to have required ribozymes whose functions were taken over by polypeptide aminoacyl-tRNA synthetases (aaRS). Two discoveries about aaRS and their tRNA substrates now furnish a unifying rationale for the opposite conclusion: that the key processes of the Central Dogma of molecular biology emerged simultaneously and naturally from simple origins in a peptide*RNA partnership, eliminating the epistemological need for a prior RNA world. First, the two aaRS classes likely arose from opposite strands of the same ancestral gene, implying a simple genetic alphabet. Inversion symmetries in aaRS structural biology arising from genetic complementarity would have stabilized the initial and subsequent differentiation of coding specificities and hence rapidly promoted diversity in the proteome. Second, amino acid physical chemistry maps onto tRNA identity elements, establishing reflexivity in protein aaRS. Bootstrapping of increasingly detailed coding is thus intrinsic to polypeptide aaRS, but impossible in an RNA world. These notions underline the following concepts that contradict gradual replacement of ribozymal aaRS by polypeptide aaRS: (i) any set of aaRS must be interdependent; (ii) reflexivity intrinsic to polypeptide aaRS production dynamics promotes bootstrapping; (iii) takeover of RNA-catalyzed aminoacylation by enzymes will necessarily degrade specificity; (iv) the Central Dogmas emergence is most probable when replication and translation error rates remain comparable. These characteristics are necessary and sufficient for the essentially de novo emergence of a coupled gene-replicase-translatase system of genetic coding that would have continuously preserved the functional meaning of genetically encoded protein genes whose phylogenetic relationships match those observed today.

evolutionary biology↗

bioRxiv2017

The Genetic Landscape Of Animal Behavior

Although most animal behaviors are associated with some form of heritable genetic variation we do not yet understand how genes sculpt behavior across evolution, either directly or indirectly. To address this, I here compile a dataset comprised of over 1,000 genomic loci representing a spectrum of behavioral variation across animal taxa. Comparative analyses reveal that courtship and feeding behaviors are associated with genomic regions of significantly greater effect than other traits, on average three fold greater than other behaviors. Investigations of whole-genome sequencing and phenotypic data for 87 behavioral traits from the Drosophila Genetics Reference Panel indicate that courtship and feeding behaviors have significantly greater genetic contributions and that, in general, behavioral traits overlap little in individual base pairs but increasingly interact at the levels of genes and traits. These results provide evidence that different types of behavior are associated with variable genetic bases and suggest that, across animal evolution, the genetic landscape of behavior is more rugged, yet predictable, than previously thought.

genomics↗

bioRxiv2017

Evolution and genetics of precocious burrowing behavior in Peromyscus mice

A central challenge in biology is to understand how innate behaviors evolve between closely related species. One way to elucidate how differences arise is to compare the development of behavior in species with distinct adult traits. Here, we report that Peromyscus polionotus is strikingly precocious with regard to burrowing behavior, but not other behaviors, compared to its sister species P. maniculatus. In P. polionotus, burrows were excavated as early as 17 days of age, while P. maniculatus did not build burrows until 10 days later. Moreover, the well-known differences in burrow architecture between adults of these species--P. polionotus adults excavate long burrows with an escape tunnel, while P. maniculatus dig short, single-tunnel burrows--were intact in juvenile burrowers. To test whether this juvenile behavior is influenced by early-life environment, pups of both species were reciprocally cross-fostered. Fostering did not alter the characteristic burrowing behavior of either species, suggesting these differences are genetic. In backcross F2 hybrids, we show that precocious burrowing and adult tunnel length are genetically correlated, and that a single P. polionotus allele in a genomic region linked to adult tunnel length is predictive of precocious burrow construction. The co-inheritance of developmental and adult traits indicates the same genetic region--either a single gene with pleiotropic effects, or closely linked genes-- acts on distinct aspects of the same behavior across life stages. Such genetic variants likely affect behavioral drive (i.e. motivation) to burrow, and thereby affect both the development and adult expression of burrowing behavior.\n\nHighlightsO_LIJuvenile P. polionotus construct burrows precociously compared to its sister species P. maniculatus\nC_LIO_LICross-fostering does not alter species-specific burrowing behavior\nC_LIO_LIA QTL linked to adult tunnel length predicts developmental onset of burrow construction in hybrids\nC_LIO_LIPleiotropic genetic variant(s) may affect behavioral drive across life stages\nC_LI

animal behavior and cognition↗

bioRxiv2017

Efficient strategies for screening large-scale genetic interaction networks

Large-scale genetic interaction screening is a powerful approach for unbiased characterization of gene function and understanding systems-level cellular organization. While genome-wide screens are desirable as they provide the most comprehensive interaction profiles, they are resource and time-intensive and sometimes infeasible, depending on the species and experimental platform. For these scenarios, optimal methods for more efficient screening while still producing the maximal amount of information from the resulting profiles are of interest.\n\nTo address this problem, we developed an optimal algorithm, called COMPRESS-GI, which selects a small but informative set of genes that captures most of the functional information contained within genome-wide genetic interaction profiles. The utility of this algorithm is demonstrated through an application of the approach to define a diagnostic mutant set for large-scale chemical genetic screens, where more than 13,000 compound screens were achieved through the increased throughput enabled by the approach. COMPRESS-GI can be broadly applied for directing genetic interaction screens in other contexts, including in species with little or no prior genetic-interaction data.

systems biology↗

bioRxiv2017

Context-Dependent Genetic Regulation

Cells process extra-cellular signals with multiple layers of complex biological networks. Due to the stochastic nature of the networks, the signals become significantly noisy within the cells and in addition, due to the nonlinear nature of the networks, the signals become distorted, shifted, and (de-)amplified. Such nonlinear signal processing can lead to non-trivial cellular phenotypes such as cell cycles, differentiation, cell-to-cell communication, and homeostasis. These nonlinear pheno-types, when observed at the cell population levels, can be quite different from the single-cell level observation. As one of the underlying mechanisms behind this difference, we report the interplay between nonlinearity and stochasticity in genetic regulation. Here we show that nonlinear genetic regulation, characterized at the cellular population level, can be affected by cell-to-cell variability in the regulatory factor concentrations. The observed genetic regulation at the cell population is shown to be significantly dependent on the upstream DNA sequences of the regulator, in particular, 5 untranslated region. This indicates that genetic regulation observed at the cell population level can be significantly dependent on its genetic context, and that its characterization needs a careful attention on noise propagation.\n\nOne Sentence SummaryGenetic regulation observed at the cell population level can be significantly affected by cell-to-cell variability in the regulatory factor copy numbers, indicating that the observed regulation is dependent on 5 UTR of the regulator coding gene.

synthetic biology↗

bioRxiv2017

The functional and genetic associations of neuroimaging data: a toolbox

Advances in neuroimaging and sequencing techniques provide an unprecedented opportunity to map the function of brain regions and to identify the roots of psychiatric diseases. However, the results generated by most neuroimaging studies, i.e., activated clusters/regions or functional connectivities between brain regions, frequently cannot be conveniently and systematically interpreted, rendering the biological meaning unclear. We describe a Brain Annotation Toolbox (BAT), a toolbox that helps to generate functional and genetic annotations for neuroimaging results. The toolbox can take data from brain regions identified with an atlas, or from brain regions identified as activated in tasks, or from functional connectivity links or networks of links. Then, the voxel-level functional description from the Neurosynth database and the gene expression profile from the Allen Brain Atlas are used to generate functional and genetic knowledge for such region-level data. Parametric (Fishers exact test) or non-parametric (permutation test) statistical tests are adopted to identify significantly related functional descriptors and genes for the neuroimaging results. The validity of the approach is demonstrated by showing that the functional and genetic annotations for specific brain regions are consistent with each other; and further the region by region functional similarity network and gene co-expression networks are highly correlated for many major brain atlases. One application of BAT is to help provide functional and genetic annotations for the newly discovered regions with unknown functions, e.g., the 97 new regions identified in the Human Connectome Project. Importantly too, this toolbox can help understand differences between patients with psychiatric disorders and controls, and this is demonstrated using data for schizophrenia and autism, for which the functional and genetic annotations for the neuroimaging data differences between patients and controls are consistent with each other and help with the interpretation of the differences.

neuroscience↗

bioRxiv2017

Identifying the genetic determinants of particular phenotypes in microbial genomes with very small training sets

Machine learning (ML) encompasses numerous algorithms that aim at discovering complex patterns between elements within large data using limited prior assumptions or modeling. However, some scientific disciplines still produce small data sets: in particular, empirical studies that try to find the mutations responsible for complex phenotypes are often limited to very small sample sizes (n), while scanning a large number of amino acid sites (p) in a proteome. To date, little is known on how ML performs in this type of so-called \"large p, small n\" problem. To address this question, we evaluated the performance of two general ML classifiers, adaptive boosting (AB) and random forest, on two data sets. To assess the impact of proteome size, we contrasted a small (viral) genome with a larger (bacterial) one. To analyze large proteomes, we further developed a chunking algorithm, and introduce a repeated random forest (RRF) algorithm that stabilizes model predictions. With the influenza data, we were able to rediscover amino acid sites experimentally implicated in three different complex phenotypes (infectivity, transmissibility, and pathogenicity). Results for the larger proteome, pertaining to three types of drug resistance (Ciprofloxacin, Ceftazidime, and Gentamicin), were more nuanced, with RRF making more sensible pre-dictions, with smaller errors rates, than AB. Furthermore, we show that chunking improved runtimes by an order of magnitude and may increase sensitivity of the predictions. Altogether, we demonstrate that ML algorithms can be used to identify genetic determinants in small proteomes (viruses), even with small numbers of individuals. We further show that even if the size of bacterial proteomes pushes AB to its limits in the context of small n, RRF may deserve more scrutiny, which should be facilitated by the plummeting costs of sequencing and, more critically, by phenotyping large cohorts of individuals.\n\nAuthor SummaryFinding the genetic determinants of a phenotype is typically performed by testing for an association between a particular allele and a trait, carrying out the testing over a large number of loci in a large cohort of individuals, itself divided into two subsets of individuals: those who have the trait (cases), and those who do not (controls). However, recruiting large cohorts can be problematic in some experimental fields, while using genotypic information rather than complete genomes can miss some mutations. To address these issues, we implemented two machine learning (ML) algorithms, tweaked for analyzing large genomes and providing stable results. The analysis of a small viral genome, for which genetic determinants of three phenotypes are already known, showed that our approach can rediscover known mutations, almost irrespective of the ML algorithm used. However, the analysis of a larger bacterial genome, for which genetic determinants of three phenotypes are unknown, suggested that the simpler of our modified algorithms performed better, returning more sensitive predictions with lower error rates. This work demonstrates the feasibility of finding genetic determinants of complex phenotypes based on a small number of complete genomes.

bioinformatics↗