Journal of Bioinformatics and Computational Biology

Cancer cells have been recently shown to express high level of short 3′UTR isoforms that can escape miRNA-mediated regulation. We present here a computational procedure for systematically identifying shortened 3′UTRs by Affymetrix 3′ microarrays. The advantage of this technology compared to more recent and promising ones such as exon arrays and RNA-Seq is that, giving the relatively small cost, already existing datasets in public databases include a considerably higher number of experiments. Moreover, the design of Affymetrix Gene Chips is well-suited for 3′UTR analysis of a large number of genes. Initially, Affymetrix individual probes are regrouped into customized probesets mapping specifically the CDS or the 3′UTR of the transcript, according to RefSeq annotation. Then, candidate 3′UTR shortening events are identified by statistical differential expression analysis of customized probesets in different biological conditions. The procedure has been applied to expression data from two ovarian adenocarcinoma datasets. Selected gene sets are significantly enriched for annotated splice variant genes as well as genes involved in estrogen dependent cancer mechanisms, confirming the validity of the proposed procedure.

Structure fluctuations and conformational changes accompany all biological processes involving macromolecules. The paper presents a classification of protein residues based on the normalized equilibrium fluctuations of the residue centers of mass in proteins and a statistical analysis of conformation changes in the side-chains upon binding. Normal mode analysis and an elastic network model were applied to a set of protein complexes to calculate the residue fluctuations and develop the residue classification. Comparison with a classification based on normalized B-factors suggests that the B-factors may underestimate protein flexibility in solvent. Our classification shows that protein loops and disordered fragments are enriched with highly fluctuating residues and depleted with weakly fluctuating residues. Strategies for engineering thermostable proteins are discussed. To calculate the dihedral angles distribution functions, the configuration space was divided into cells by a cubic grid. The effect of protein association on the distribution functions depends on the amino acid type and a grid step in the dihedral angles space. The changes in the dihedral angles increase from the near-backbone dihedral angle to the most distant one, for most residues. On average, one fifth of the interface residues change the rotamer state upon binding, whereas the rest of the interface residues undergo local readjustments within the same rotamer.

The description of a protein fold is a hard problem due to significant variability of main structural units, β-sheets and α-helixes, and their mutual arrangements. An adequate description of the structural units is an important step in objective protein structure classification, which to date is based on expert judgment in a number of cases. Explicit determination and description of structural units is more complicated for β-sheets than for α-helixes due to β-sheets variability both in composition and geometry. We have developed an algorithm that can significantly modify β-sheets detected by commonly used DSSP and Stride algorithms and represent the result as a “β-sheet map,” a table describing certain β-sheet features. In our approach, β-sheets (rather than β-strands) are considered as holistic objects. Both hydrogen bonds and geometrical restrains are explored for the determination of β-sheets. The algorithm is implemented in SheeP program. It was tested for prediction architectures of domains from 93 well-defined all-β and α/β SCOP protein domain families, and showed 93% of correct results. The Web-service http://mouse.belozersky.msu.ru/sheep allows to detect β-sheets in a given protein structure, visualize β-sheet maps, as well as input three-dimensional structures with highlighted β-sheets and their structural features.

Electrostatic properties of genome DNA are important to its interactions with different proteins, in particular, related to transcription. DEPPDB — DNA Electrostatic Potential (and other Physical) Properties Database — provides information on the electrostatic and other physical properties of genome DNA combined with its sequence and annotation of biological and structural properties of genomes and their elements. Genomes are organized on taxonomical basis, supporting comparative and evolutionary studies. Currently, DEPPDB contains all completely sequenced bacterial, viral, mitochondrial, and plastids genomes according to the NCBI RefSeq, and some model eukaryotic genomes. Data for promoters, regulation sites, binding proteins, etc., are incorporated from established DBs and literature. The database is complemented by analytical tools. User sequences calculations are available. Case studies discovered electrostatics complementing DNA bending in E.coli plasmid BNT2 promoter functioning, possibly affecting host-environment metabolic switch. Transcription factors binding sites gravitate to high potential regions, confirming the electrostatics universal importance in protein–DNA interactions beyond the classical promoter–RNA polymerase recognition and regulation. Other genome elements, such as terminators, also show electrostatic peculiarities. Most intriguing are gene starts, exhibiting taxonomic correlations. The necessity of the genome electrostatic properties studies is discussed.

The new generation of short-read sequencing technologies requires reliable measures of data quality. Such measures are especially important for variant calling. However, in the particular case of SNP calling, a great number of false-positive SNPs may be obtained. One needs to distinguish putative SNPs from sequencing or other errors. We found that not only the probability of sequencing errors (i.e. the quality value) is important to distinguish an FP-SNP but also the conditional probability of "correcting" this error (the "second best call" probability, conditional on that of the first call). Surprisingly, around 80% of mismatches can be "corrected" with this second call. Another way to reduce the rate of FP-SNPs is to retrieve DNA motifs that seem to be prone to sequencing errors, and to attach a corresponding conditional quality value to these motifs. We have developed several measures to distinguish between sequence errors and candidate SNPs, based on a base call's nucleotide context and its mismatch type. In addition, we suggested a simple method to correct the majority of mismatches, based on conditional probability of their "second" best intensity call. We attach a corresponding second call confidence (quality value) of being corrected to each mismatch.

Commonly used RNA folding programs compute the minimum free energy structure of a sequence under the pseudoknot exclusion constraint. They are based on Zuker's algorithm which runs in time O(n³). Recently, it has been claimed that RNA folding can be achieved in average time O(n²) using a sparsification technique. A proof of quadratic time complexity was based on the assumption that computational RNA folding obeys the "polymer-zeta property". Several variants of sparse RNA folding algorithms were later developed. Here, we present our own version, which is readily applicable to existing RNA folding programs, as it is extremely simple and does not require any new data structure. We applied it to the widely used Vienna RNAfold program, to create sibRNAfold, the first public sparsified version of a standard RNA folding program. To gain a better understanding of the time complexity of sparsified RNA folding in general, we carried out a thorough run time analysis with synthetic random sequences, both in the context of energy minimization and base pairing maximization. Contrary to previous claims, the asymptotic time complexity of a sparsified RNA folding algorithm using standard energy parameters remains O(n³) under a wide variety of conditions. Consistent with our run-time analysis, we found that RNA folding does not obey the "polymer-zeta property" as claimed previously. Yet, a basic version of a sparsified RNA folding algorithm provides 15- to 50-fold speed gain. Surprisingly, the same sparsification technique has a different effect when applied to base pairing optimization. There, its asymptotic running time complexity appears to be either quadratic or cubic depending on the base composition. The code used in this work is available at: http://sibRNAfold.sourceforge.net/.

Evolutionary aspects of the genetic architecture of common human diseases remain enigmatic. The results of more than 200 genome-wide association studies published to date were compiled in a catalog (http://www.genome.gov/26525384/). We used cataloged data to determine whether derived (mutant) alleles are associated with higher risk of human disease more frequently than ancestral alleles. We placed all allelic variants into ten categories of population frequency (0%–100%) in 10% increments. We then analyzed the relationship between allelic frequency, evolutionary status of the polymorphic site (ancestral versus derived), and disease risk status (risk versus protection). Given the same population frequency, derived alleles are more likely to be risk associated than ancestral alleles, as are rarer alleles. The common interpretation of this association is that negative selection prevents fixation of the risk variants. However, disease stratification as early or late onset suggests that weak selection against risk-associated alleles is unlikely a major factor shaping genetic architecture of common diseases. Our results clearly suggest that the duration of existence of an allele in a population is more important. Alleles existing longer tend to show weaker linkage disequilibrium with neighboring alleles, including the causal alleles, and are less likely to tag a SNP-disease association.

The transfer of single-stranded (ss) T-DNA from soil bacteria of the genus Agrobacterium with the help of the VirE2 protein, which possibly mediates the delivery of ss-T-DNA across the cell membrane, was demonstrated earlier, but how VirE2 participates in ssDNA transfer across artificial and natural membranes is not known. Using computational methods, we reconstructed model structures composed of two and four VirE2 proteins and showed by the MOLE program the formation of pores with channel diameters of 1.2–1.6 and 1.4–4.6 nm in a model structure formed from two and four VirE2 molecules, respectively. Using light scattering, we recorded the size distribution for recombinant VirE2-dependent complexes in aqueous solutions and found that VirE2 in a buffer solution is present as a complex made up of two or more proteins. We revealed single, long-lived jumps in voltage-dependent membrane conductance during coincubation of planar black membranes with the VirE2 protein. On the addition of VirE2 and FAM-labeled oligonucleotides to HeLa cells, the fluorescence intensity for the cells increased by 56% as compared to that for cells incubated only with oligonucleotides.

Ribonucleic acid (RNA) molecules play important roles in a variety of biological processes. To properly function, RNA molecules usually have to fold to specific structures, and therefore understanding RNA structure is vital in comprehending how RNA functions. One approach to understanding and predicting biomolecular structure is to use knowledge-based potentials built from experimentally determined structures. These types of potentials have been shown to be effective for predicting both protein and RNA structures, but their utility is limited by their significantly rugged nature. This ruggedness (and hence the potential's usefulness) depends heavily on the choice of bin width to sort structural information (e.g. distances) but the appropriate bin width is not known a priori. To circumvent the binning problem, we compared knowledge-based potentials built from inter-atomic distances in RNA structures using different mixture models (Kernel Density Estimation, Expectation Minimization and Dirichlet Process). We show that the smooth knowledge-based potential built from Dirichlet process is successful in selecting native-like RNA models from different sets of structural decoys with comparable efficacy to a potential developed by spline-fitting — a commonly taken approach — to binned distance histograms. The less rugged nature of our potential suggests its applicability in diverse types of structural modeling.

We consider networks with two types of nodes. The v-nodes, called centers, are hyperconnected and interact with one another via many u-nodes, called satellites. This centralized architecture, widespread in gene networks, possesses two fundamental properties. Namely, this organization creates feedback loops that are capable of generating practically any prescribed patterning dynamics, chaotic or periodic, or having a number of equilibrium states. Moreover, this organization is robust with respect to random perturbations of the system.

It has been long experimentally demonstrated that human alpha-fetoprotein (HAFP) has an ability to bind immobilized estrogens with the most efficiency for synthetic estrogen analog — diethylstilbestrol (DES). However, the question remains why the human AFP (HAFP), unlike rodent AFP, cannot bind free estrogens. Moreover, despite the fact that AFP was first discovered more than 50 years ago and is presently recognized as a "golden standard" among onco-biomarkers, its three-dimensional (3D) structure has not been experimentally solved yet. In this work using MODELLER program, we generated 3D model of HAFP on the basis of homology with human serum albumin (HSA) and Vitamin D–binding protein (VTDB) with subsequent molecular docking of DES to the model structure and molecular dynamics (MD) simulation study of the complex obtained. The model constructed has U-shaped structure in which a cavity may be distinguished. In this cavity the putative estrogen-binding site is localized. Validation by RMSD calculation and with the use of PROCHECK program showed good quality of the model and stability of extended region of four alpha-helical structures that contains putative hormone-binding residues. Data extracted from MD simulation trajectory allow proposing two types of interactions between amino acid residues of HAFP and DES molecule: (1) hydrogen bonding with involvement of residues S445, R452, and E551; (2) hydrophobic interactions with participation of L138, M448, and M548 residues. A suggestion is made that immobilization of the hormone using a long spacer provides delivery of the estrogen molecule to the binding site and, thereby, facilitates interaction between HAFP and the hormone.

Identifying genes associated with cancer development is typically accomplished by comparing mean expression values in normal and tumor tissues, which identifies differentially expressed (DE) genes. Interindividual variation (IV) in gene expression is indirectly included in DE gene identification because given the same absolute differences in means, genes with lower variance tend to have lower p-values. We explored the direct use of IV in gene expression to identify candidate genes associated with cancer development. We focused on prostate (PCa) and lung (LC) cancers and compared IV in the expression level of genes shown to be cancer related with that in all other genes in the human genome. Compared with all those other genes, cancer-related genes tended to have greater IV in normal tissues and a greater increase in IV during the transition from normal to tumorous tissue. Genes without significantly different mean expression values between tumor and normal tissues but with greater IV in tumor than in normal tissue (note: the DE-based approach completely ignores those genes) had stronger associations with clinically important features like Gleason score in PCa or tumor histology in LC than all other genes were. Our results suggest that analyzing IV in gene expression level is useful in identifying novel candidate genes associated with cancer development.