Alexander V. Favorov

A Gibbs sampler for identification of symmetrically structured, spaced DNA motifs with improved estimation of the signal length

MOTIVATION Transcription regulatory protein factors often bind DNA as homo-dimers or hetero-dimers. Thus they recognize structured DNA motifs that are inverted or direct repeats or spaced motif pairs. However, these motifs are often difficult to identify owing to their high divergence. The motif structure included explicitly into the motif recognition algorithm improves recognition efficiency for highly divergent motifs as well as estimation of motif geometric parameters. RESULT We present a modification of the Gibbs sampling motif extraction algorithm, SeSiMCMC (Sequence Similarities by Markov Chain Monte Carlo), which finds structured motifs of these types, as well as non-structured motifs, in a set of unaligned DNA sequences. It employs improved estimators of motif and spacer lengths. The probability that a sequence does not contain any motif is accounted for in a rigorous Bayesian manner. We have applied the algorithm to a set of upstream regions of genes from two Escherichia coli regulons involved in respiration. We have demonstrated that accounting for a symmetric motif structure allows the algorithm to identify weak motifs more accurately. In the examples studied, ArcA binding sites were demonstrated to have the structure of a direct spaced repeat, whereas NarP binding sites exhibited the palindromic structure. AVAILABILITY The WWW interface of the program, its FreeBSD (4.0) and Windows 32 console executables are available at http://bioinform.genetika.ru/SeSiMCMC

Exploring Massive, Genome Scale Datasets with the GenometriCorr Package

We have created a statistically grounded tool for determining the correlation of genomewide data with other datasets or known biological features, intended to guide biological exploration of high-dimensional datasets, rather than providing immediate answers. The software enables several biologically motivated approaches to these data and here we describe the rationale and implementation for each approach. Our models and statistics are implemented in an R package that efficiently calculates the spatial correlation between two sets of genomic intervals (data and/or annotated features), for use as a metric of functional interaction. The software handles any type of pointwise or interval data and instead of running analyses with predefined metrics, it computes the significance and direction of several types of spatial association; this is intended to suggest potentially relevant relationships between the datasets. Availability and implementation: The package, GenometriCorr, can be freely downloaded at http://genometricorr.sourceforge.net/. Installation guidelines and examples are available from the sourceforge repository. The package is pending submission to Bioconductor.

Inferring causal molecular networks: empirical assessment through a community-based effort

It remains unclear whether causal, rather than merely correlational, relationships in molecular networks can be inferred in complex biological settings. Here we describe the HPN-DREAM network inference challenge, which focused on learning causal influences in signaling networks. We used phosphoprotein data from cancer cell lines as well as in silico data from a nonlinear dynamical model. Using the phosphoprotein data, we scored more than 2,000 networks submitted by challenge participants. The networks spanned 32 biological contexts and were scored in terms of causal validity with respect to unseen interventional data. A number of approaches were effective, and incorporating known biology was generally advantageous. Additional sub-challenges considered time-course prediction and visualization. Our results suggest that learning causal relationships may be feasible in complex settings such as disease states. Furthermore, our scoring approach provides a practical way to empirically assess inferred molecular networks in a causal sense.

RNAKinetics: a web server that models secondary structure kinetics of an elongating RNA.

The RNAKinetics server (http://www.ig-msk.ru/RNA/kinetics) is a web interface for the newly developed RNAKinetics software. The software models the dynamics of RNA secondary structure by the means of kinetic analysis of folding transitions of a growing RNA molecule. The result of the modeling is a kinetic ensemble, i.e. a collection of RNA structures that are endowed with probabilities, which depend on time. This approach gives comprehensive probabilistic description of RNA folding pathways, revealing important kinetic details that are not captured by the traditional structure prediction methods. The access to the RNAKinetics server is free.

A Markov chain Monte Carlo technique for identification of combinations of allelic variants underlying complex diseases in humans.

In recent years, the number of studies focusing on the genetic basis of common disorders with a complex mode of inheritance, in which multiple genes of small effect are involved, has been steadily increasing. An improved methodology to identify the cumulative contribution of several polymorphous genes would accelerate our understanding of their importance in disease susceptibility and our ability to develop new treatments. A critical bottleneck is the inability of standard statistical approaches, developed for relatively modest predictor sets, to achieve power in the face of the enormous growth in our knowledge of genomics. The inability is due to the combinatorial complexity arising in searches for multiple interacting genes. Similar “curse of dimensionality” problems have arisen in other fields, and Bayesian statistical approaches coupled to Markov chain Monte Carlo (MCMC) techniques have led to significant improvements in understanding. We present here an algorithm, APSampler, for the exploration of potential combinations of allelic variations positively or negatively associated with a disease or with a phenotype. The algorithm relies on the rank comparison of phenotype for individuals with and without specific patterns (i.e., combinations of allelic variants) isolated in genetic backgrounds matched for the remaining significant patterns. It constructs a Markov chain to sample only potentially significant variants, minimizing the potential of large data sets to overwhelm the search. We tested APSampler on a simulated data set and on a case-control MS (multiple sclerosis) study for ethnic Russians. For the simulated data, the algorithm identified all the phenotype-associated allele combinations coded into the data and, for the MS data, it replicated the previously known findings.

Genetic polymorphisms, their allele combinations and IFN-β treatment response in Irish multiple sclerosis patients

INTRODUCTION IFN-beta is widely used as first-line immunomodulatory treatment for multiple sclerosis. Response to treatment is variable (30-50% of patients are nonresponders) and requires a long treatment duration for accurate assessment to be possible. Information about genetic variations that predict responsiveness would allow appropriate treatment selection early after diagnosis, improve patient care, with time saving consequences and more efficient use of resources. MATERIALS & METHODS We analyzed 61 SNPs in 34 candidate genes as possible determinants of IFN-beta response in Irish multiple sclerosis patients. Particular emphasis was placed on the exploration of combinations of allelic variants associated with response to therapy by means of a Markov chain Monte Carlo-based approach (APSampler). RESULTS The most significant allelic combinations, which differed in frequency between responders and nonresponders, included JAK2-IL10RB-GBP1-PIAS1 (permutation p-value was p(perm) = 0.0008), followed by JAK2-IL10-CASP3 (p(perm) = 0.001). DISCUSSION The genetic mechanism of response to IFN-beta is complex and as yet poorly understood. Data mining algorithms may help in uncovering hidden allele combinations involved in drug response versus nonresponse.

Conserved and species-specific alternative splicing in mammalian genomes

BackgroundAlternative splicing has been shown to be one of the major evolutionary mechanisms for protein diversification and proteome expansion, since a considerable fraction of alternative splicing events appears to be species- or lineage-specific. However, most studies were restricted to the analysis of cassette exons in pairs of genomes and did not analyze functionality of the alternative variants.ResultsWe analyzed conservation of human alternative splice sites and cassette exons in the mouse and dog genomes. Alternative exons, especially minor-isofom ones, were shown to be less conserved than constitutive exons. Frame-shifting alternatives in the protein-coding regions are less conserved than frame-preserving ones. Similarly, the conservation of alternative sites is highest for evenly used alternatives, and higher when the distance between the sites is divisible by three. The rate of alternative-exon and site loss in mouse is slightly higher than in dog, consistent with faster evolution of the former. The evolutionary dynamics of alternative sites was shown to be consistent with the model of random activation of cryptic sites.ConclusionConsistent with other studies, our results show that minor cassette exons are less conserved than major-alternative and constitutive exons. However, our study provides evidence that this is caused not only by exon birth, but also lineage-specific loss of alternative exons and sites, and it depends on exon functionality.

Three allele combinations associated with Multiple Sclerosis

BackgroundMultiple sclerosis (MS) is an immune-mediated disease of polygenic etiology. Dissection of its genetic background is a complex problem, because of the combinatorial possibilities of gene-gene interactions. As genotyping methods improve throughput, approaches that can explore multigene interactions appropriately should lead to improved understanding of MS.Methods286 unrelated patients with definite MS and 362 unrelated healthy controls of Russian descent were genotyped at polymorphic loci (including SNPs, repeat polymorphisms, and an insertion/deletion) of the DRB1, TNF, LT, TGFβ1, CCR5 and CTLA4 genes and TNFa and TNFb microsatellites. Each allele carriership in patients and controls was compared by Fishers exact test, and disease-associated combinations of alleles in the data set were sought using a Bayesian Markov chain Monte Carlo-based method recently developed by our group.ResultsWe identified two previously unknown MS-associated tri-allelic combinations:-509TGFβ1*C, DRB1*18(3), CTLA4*G and -238TNF*B1,-308TNF*A2, CTLA4*G, which perfectly separate MS cases from controls, at least in the present sample. The previously described DRB1*15(2) allele, the microsatellite TNFa9 allele and the biallelic combination CCR5Δ32, DRB1*04 were also reidentified as MS-associated.ConclusionThese results represent an independent validation of MS association with DRB1*15(2) and TNFa9 in Russians and are the first to find the interplay of three loci in conferring susceptibility to MS. They demonstrate the efficacy of our approach for the identification of complex-disease-associated combinations of alleles.

Allelic combinations of immune-response genes associated with glatiramer acetate treatment response in Russian multiple sclerosis patients

BACKGROUND Glatiramer acetate (GA) is widely used as a first-line disease-modifying treatment for multiple sclerosis (MS). However, a significant proportion of MS patient appears to experience modest benefit from GA-treatment. Genetic variants affecting the clinical response to GA are believed to be relevant as biomarkers of GA-treatment efficiency. PATIENTS & METHODS Nine polymorphisms in candidate genes were analyzed as possible determinants of GA response in 285 Russian MS patients. Special attention was given to identification of response-associated allelic combinations by means of the APSampler algorithm. RESULTS No significant associations were found for individual polymorphisms. Alleles DRB1*15, TGFB1*T, CCR5*d and IFNAR1*G were the components of the combinations, of which carriage was significantly higher in nonresponders than in responders. Carriers of the most significant combinations: DRB1*15 + TGFB1*T + CCR5*d + IFNAR1*G and DRB1*15 + TGFB1*T + CCR5*d (permutation p-values: 0.0056 and 0.013, respectively) had a 14 to 15-times increased risk of ineffective response to GA therapy. DISCUSSION The results suggest that the influence of immune-response genes on GA-induced response has a polygenic nature. The data are interpreted as evidence of additive and epistatic influences of the genes on GA efficiency for MS treatment.

Heavy–light chain interrelations of MS-associated immunoglobulins probed by deep sequencing and rational variation ☆

The mechanisms triggering most of autoimmune diseases are still obscure. Autoreactive B cells play a crucial role in the development of such pathologies and, in particular, production of autoantibodies of different specificities. The combination of deep-sequencing technology with functional studies of antibodies selected from highly representative immunoglobulin combinatorial libraries may provide unique information on specific features in the repertoires of autoreactive B cells. Here, we have analyzed cross-combinations of the variable regions of human immunoglobulins against the myelin basic protein (MBP) previously selected from a multiple sclerosis (MS)-related scFv phage-display library. On the other hand, we have performed deep sequencing of the sublibraries of scFvs against MBP, Epstein-Barr virus (EBV) latent membrane protein 1 (LMP1), and myelin oligodendrocyte glycoprotein (MOG). Bioinformatics analysis of sequencing data and surface plasmon resonance (SPR) studies have shown that it is the variable fragments of antibody heavy chains that mainly determine both the affinity of antibodies to the parent autoantigen and their cross-reactivity. It is suggested that LMP1-cross-reactive anti-myelin autoantibodies contain heavy chains encoded by certain germline gene segments, which may be a hallmark of the EBV-specific B cell subpopulation involved in MS triggering.