Perturbation of Interaction Networks for Application to Cancer Therapy

We present a computational approach for studying the effect of potential drug combinations on the protein networks associated with tumor cells. The majority of therapeutics are designed to target single proteins, yet most diseased states are characterized by a combination of many interacting genes and proteins. Using the topology of protein-protein interaction networks, our methods can explicitly model the possible synergistic effect of targeting multiple proteins using drug combinations in different cancer types.

The methodology can be conceptually split into two distinct stages. Firstly, we integrate protein interaction and gene expression data to develop network representations of different tissue types and cancer types. Secondly, we model network perturbations to search for target combinations which cause significant damage to a relevant cancer network but only minimal damage to an equivalent normal network. We have developed sets of predicted target and drug combinations for multiple cancer types, which are validated using known cancer and drug associations, and are currently in experimental testing for prostate cancer. Our methods also revealed significant bias in curated interaction data sources towards targets with associations compared with high-throughput data sources from model organisms. The approach developed can potentially be applied to many other diseased cell types.

Abstract PD01-08: Decoding the Transcriptional Landscape of Triple-Negative Breast Cancer Using Next-Generation Whole Transcriptome Sequencing

Background: Triple-negative breast cancer (TNBC) has been plagued by the absence of targeted therapies. Discovery of therapeutic targets in TNBC has in part, been hampered by an inadequate understanding of the transcriptional biology of the normal breast as an optimal comparator. Using next-generation sequencing, we embarked on a study to compare the transcriptomes of TNBC and normal breast to comprehensively identify novel targets by analyzing all full length transcripts expressed in these tissues.

Methods: Normal breast tissues from healthy pre-menopausal volunteers with no history of disease were procured from the Susan G. Komen for the Cure® Tissue Bank at the IU Simon Cancer Center. To eliminate bias from stromal tissue, normal tissues were laser capture microdissected for ductal epithelium. cDNA libraries from 10 TNBC tumors and 10 normal breast tissues were sequenced on an Applied Biosystems (AB) SOLiD3 sequencer using 50bp fragment runs. For gene expression, mapping of reads to the genome was performed using the AB BioScope 1.2 Pipeline and outputs imported into Partek Genomics Suite for analysis. In Partek, mapped reads were cross-referenced against known genes from the UCSC database followed by statistical comparison of RPKM values for each gene between TNBC and normal. Dimensionality reduction analyses (PCA & Hierarchical clustering) and identification of Novel Transcribed Regions were also performed in Partek, whereas construction of gene networks was performed using Ingenuity Pathway Analysis. To identify gene fusions, partially mapped reads were interrogated utilizing a novel algorithm that searched for reads spanning exons from two different genes. Fusions that were supported by at least 3 reads (of which 2 had to be unique) were considered candidates and were subsequently validated. Results/Discussion: Sequencing produced 1.1 billion reads equaling 57.3GB of data of which 36.0GB (63%) mapped to the human genome. In comparing RPKM values between TNBC and Normal, we report 7140 RefSeq Genes, 22 pre-miRNAs, 109 lincRNA exons, and 15 ultraconserved regions that were differentially expressed between these tissues (FDR<0.01). Biological interpretation of these results reveals upregulation of genes and miRNAs involved in DNA repair, angiogenesis, and inhibitors of Estrogen Receptor-alpha. Some previous drug targets (e.g. EGFR and c-kit) were not found to be upregulated here which may explain lack of clinical success to date. Conversely, PARP was significantly upregulated and early trial results suggest a strong signal for efficacy with inhibition of PARP. We also surveyed the genome for Novel Transcribed Regions (NTRs), defined as areas of significant transcription where no annotated gene is present. When comparing between TNBC and Normal, we report 6408 NTRs to be differentially expressed (FDR<0.01). Lastly, when analyzing the dataset for gene fusions, we identified several gene fusions in the TNBC samples, though no individual fusion was present in more than one sample.

Conclusion: We report an extensive comparison of the transcriptomes of TNBC and normal ductal epithelium. We identified numerous genes previously unknown to be dysregulated in TNBC that can be utilized for therapeutic discovery.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr PD01-08.

Perturbation of interaction networks for application to cancer therapy

We present a computational approach for studying the effect of potential drug combinations on the protein networks associated with tumor cells. The majority of therapeutics are designed to target single proteins, yet most diseased states are characterized by a combination of many interacting genes and proteins. Using the topology of protein-protein interaction networks, our methods can explicitly model the possible synergistic effect of targeting multiple proteins using drug combinations in different cancer types. The methodology can be conceptually split into two distinct stages. Firstly, we integrate protein interaction and gene expression data to develop network representations of different tissue types and cancer types. Secondly, we model network perturbations to search for target combinations which cause significant damage to a relevant cancer network but only minimal damage to an equivalent normal network. We have developed sets of predicted target and drug combinations for multiple cancer types, which are validated using known cancer and drug associations, and are currently in experimental testing for prostate cancer. Our methods also revealed significant bias in curated interaction data sources towards targets with associations compared with high-throughput data sources from model organisms. The approach developed can potentially be applied to many other diseased cell types.

The complete genome of Rhodococcus sp. RHA1 provides insights into a catabolic powerhouse

Rhodococcus sp. RHA1 (RHA1) is a potent polychlorinated biphenyl-degrading soil actinomycete that catabolizes a wide range of compounds and represents a genus of considerable industrial interest. RHA1 has one of the largest bacterial genomes sequenced to date, comprising 9,702,737 bp (67% G+C) arranged in a linear chromosome and three linear plasmids. A targeted insertion methodology was developed to determine the telomeric sequences. RHA1's 9,145 predicted protein-encoding genes are exceptionally rich in oxygenases (203) and ligases (192). Many of the oxygenases occur in the numerous pathways predicted to degrade aromatic compounds (30) or steroids (4). RHA1 also contains 24 nonribosomal peptide synthase genes, six of which exceed 25 kbp, and seven polyketide synthase genes, providing evidence that rhodococci harbor an extensive secondary metabolism. Among sequenced genomes, RHA1 is most similar to those of nocardial and mycobacterial strains. The genome contains few recent gene duplications. Moreover, three different analyses indicate that RHA1 has acquired fewer genes by recent horizontal transfer than most bacteria characterized to date and far fewer than Burkholderia xenovorans LB400, whose genome size and catabolic versatility rival those of RHA1. RHA1 and LB400 thus appear to demonstrate that ecologically similar bacteria can evolve large genomes by different means. Overall, RHA1 appears to have evolved to simultaneously catabolize a diverse range of plant-derived compounds in an O(2)-rich environment. In addition to establishing RHA1 as an important model for studying actinomycete physiology, this study provides critical insights that facilitate the exploitation of these industrially important microorganisms.

Analysis of the prostate cancer cell line LNCaP transcriptome using a sequencing-by-synthesis approach

Background

High throughput sequencing-by-synthesis is an emerging technology that allows the rapid production of millions of bases of data. Although the sequence reads are short, they can readily be used for re-sequencing. By re-sequencing the mRNA products of a cell, one may rapidly discover polymorphisms and splice variants particular to that cell.

Results

We present the utility of massively parallel sequencing by synthesis for profiling the transcriptome of a human prostate cancer cell-line, LNCaP, that has been treated with the synthetic androgen, R1881. Through the generation of approximately 20 megabases (MB) of EST data, we detect transcription from over 10,000 gene loci, 25 previously undescribed alternative splicing events involving known exons, and over 1,500 high quality single nucleotide discrepancies with the reference human sequence. Further, we map nearly 10,000 ESTs to positions on the genome where no transcription is currently predicted to occur. We also characterize various obstacles with using sequencing by synthesis for transcriptome analysis and propose solutions to these problems.

Conclusion

The use of high-throughput sequencing-by-synthesis methods for transcript profiling allows the specific and sensitive detection of many of a cell's transcripts, and also allows the discovery of high quality base discrepancies, and alternative splice variants. Thus, this technology may provide an effective means of understanding various disease states, discovering novel targets for disease treatment, and discovery of novel transcripts.

Sequence biases in large scale gene expression profiling data

We present the results of a simple, statistical assay that measures the G+C content sensitivity bias of gene expression experiments without the requirement of a duplicate experiment. We analyse five gene expression profiling methods: Affymetrix GeneChip, Long Serial Analysis of Gene Expression (LongSAGE), LongSAGELite, 'Classic' Massively Parallel Signature Sequencing (MPSS) and 'Signature' MPSS. We demonstrate the methods have systematic and random errors leading to a different G+C content sensitivity. The relationship between this experimental error and the G+C content of the probe set or tag that identifies each gene influences whether the gene is detected and, if detected, the level of gene expression measured. LongSAGE has the least bias, while Signature MPSS shows a strong bias to G+C rich tags and Affymetrix data show different bias depending on the data processing method (MAS 5.0, RMA or GC-RMA). The bias in the Affymetrix data primarily impacts genes expressed at lower levels. Despite the larger sampling of the MPSS library, SAGE identifies significantly more genes (60% more RefSeq genes in a single comparison).

Physical map-assisted whole-genome shotgun sequence assemblies

We describe a targeted approach to improve the contiguity of whole-genome shotgun sequence (WGS) assemblies at run-time, using information from Bacterial Artificial Chromosome (BAC)-based physical maps. Clone sizes and overlaps derived from clone fingerprints are used for the calculation of length constraints between any two BAC neighbors sharing 40% of their size. These constraints are used to promote the linkage and guide the arrangement of sequence contigs within a sequence scaffold at the layout phase of WGS assemblies. This process is facilitated by FASSI, a stand-alone application that calculates BAC end and BAC overlap length constraints from clone fingerprint map contigs created by the FPC package. FASSI is designed to work with the assembly tool PCAP, but its output can be formatted to work with other WGS assembly algorithms able to use length constraints for individual clones. The FASSI method is simple to implement, potentially cost-effective, and has resulted in the increase of scaffold contiguity for both the Drosophila melanogaster and Cryptococcus gattii genomes when compared to a control assembly without map-derived constraints. A 6.5-fold coverage draft DNA sequence of the Pan troglodytes (chimpanzee) genome was assembled using map-derived constraints and resulted in a 26.1% increase in scaffold contiguity.

Sequencing and analysis of 10,967 full-length cDNA clones from Xenopus laevis and Xenopus tropicalis reveals post-tetraploidization transcriptome remodeling

Sequencing of full-insert clones from full-length cDNA libraries from both Xenopus laevis and Xenopus tropicalis has been ongoing as part of the Xenopus Gene Collection Initiative. Here we present 10,967 full ORF verified cDNA clones (8049 from X. laevis and 2918 from X. tropicalis) as a community resource. Because the genome of X. laevis, but not X. tropicalis, has undergone allotetraploidization, comparison of coding sequences from these two clawed (pipid) frogs provides a unique angle for exploring the molecular evolution of duplicate genes. Within our clone set, we have identified 445 gene trios, each comprised of an allotetraploidization-derived X. laevis gene pair and their shared X. tropicalis ortholog. Pairwise dN/dS, comparisons within trios show strong evidence for purifying selection acting on all three members. However, dN/dS ratios between X. laevis gene pairs are elevated relative to their X. tropicalis ortholog. This difference is highly significant and indicates an overall relaxation of selective pressures on duplicated gene pairs. We have found that the paralogs that have been lost since the tetraploidization event are enriched for several molecular functions, but have found no such enrichment in the extant paralogs. Approximately 14% of the paralogous pairs analyzed here also show differential expression indicative of subfunctionalization.

Assessment and integration of publicly available SAGE, cDNA microarray, and oligonucleotide microarray expression data for global coexpression analyses

Large amounts of gene expression data from several different technologies are becoming available to the scientific community. A common practice is to use these data to calculate global gene coexpression for validation or integration of other "omic" data. To assess the utility of publicly available datasets for this purpose we have analyzed Homo sapiens data from 1202 cDNA microarray experiments, 242 SAGE libraries, and 667 Affymetrix oligonucleotide microarray experiments. The three datasets compared demonstrate significant but low levels of global concordance (rc<0.11). Assessment against Gene Ontology (GO) revealed that all three platforms identify more coexpressed gene pairs with common biological processes than expected by chance. As the Pearson correlation for a gene pair increased it was more likely to be confirmed by GO. The Affymetrix dataset performed best individually with gene pairs of correlation 0.9-1.0 confirmed by GO in 74% of cases. However, in all cases, gene pairs confirmed by multiple platforms were more likely to be confirmed by GO. We show that combining results from different expression platforms increases reliability of coexpression. A comparison with other recently published coexpression studies found similar results in terms of performance against GO but with each method producing distinctly different gene pair lists.

cisRED: a database system for genome-scale computational discovery of regulatory elements

We describe cisRED, a database for conserved regulatory elements that are identified and ranked by a genome-scale computational system (www.cisred.org). The database and high-throughput predictive pipeline are designed to address diverse target genomes in the context of rapidly evolving data resources and tools. Motifs are predicted in promoter regions using multiple discovery methods applied to sequence sets that include corresponding sequence regions from vertebrates. We estimate motif significance by applying discovery and post-processing methods to randomized sequence sets that are adaptively derived from target sequence sets, retain motifs with p-values below a threshold and identify groups of similar motifs and co-occurring motif patterns. The database offers information on atomic motifs, motif groups and patterns. It is web-accessible, and can be queried directly, downloaded or installed locally.

A mouse atlas of gene expression: large-scale digital gene-expression profiles from precisely defined developing C57BL/6J mouse tissues and cells

We analyzed 8.55 million LongSAGE tags generated from 72 libraries. Each LongSAGE library was prepared from a different mouse tissue. Analysis of the data revealed extensive overlap with existing gene data sets and evidence for the existence of approximately 24,000 previously undescribed genomic loci. The visual cortex, pancreas, mammary gland, preimplantation embryo, and placenta contain the largest number of differentially expressed transcripts, 25% of which are previously undescribed loci.

Management and visualization of whole genome shotgun assemblies using SAM

We have designed and implemented a system to manage whole genome shotgun sequences and whole genome sequence assembly data flow. The Sequence Assembly Manager (SAM) consists primarily of a MySQL relational database and Perl applications designed to easily manipulate and coordinate the analysis of sequence information and to view and report genome assembly progress through its Common Gateway Interface (CGI) web interface. The application includes a tool to compare sequence assemblies to fingerprint maps that has been used successfully to improve and validate both maps and sequence assemblies of the Rhodococcus sp.RHAI and Cryptococcus neoformans WM276 genomes.

Sockeye: a 3D environment for comparative genomics

Comparative genomics techniques are used in bioinformatics analyses to identify the structural and functional properties of DNA sequences. As the amount of available sequence data steadily increases, the ability to perform large-scale comparative analyses has become increasingly relevant. In addition, the growing complexity of genomic feature annotation means that new approaches to genomic visualization need to be explored. We have developed a Java-based application called Sockeye that uses three-dimensional (3D) graphics technology to facilitate the visualization of annotation and conservation across multiple sequences. This software uses the Ensembl database project to import sequence and annotation information from several eukaryotic species. A user can additionally import their own custom sequence and annotation data. Individual annotation objects are displayed in Sockeye by using custom 3D models. Ensembl-derived and imported sequences can be analyzed by using a suite of multiple and pair-wise alignment algorithms. The results of these comparative analyses are also displayed in the 3D environment of Sockeye. By using the Java3D API to visualize genomic data in a 3D environment, we are able to compactly display cross-sequence comparisons. This provides the user with a novel platform for visualizing and comparing genomic feature organization.

Protein structural domains: analysis of the 3Dee domains database

The 3Dee database of domain definitions was developed as a comprehensive collection of domain definitions for all three-dimensional structures in the Protein Data Bank (PDB). The database includes definitions for complex, multiple-segment and multiple-chain domains as well as simple sequential domains, organized in a structural hierarchy. Two different snapshots of the 3Dee database were analyzed at September 1996 and November 1999. For the November 1999 release, 7,995 PDB entries contained 13,767 protein chains and gave rise to 18,896 domains. The domain sequences clustered into 1,715 domain sequence families, which were further clustered into a conservative 1,199 domain structure families (families with similar folds). The proportion of different domain structure families per domain sequence family increases from 84% for domains 1-100 residues long to 100% for domains greater than 600 residues. This is in keeping with the idea that longer chains will have more alternative folds available to them. Of the representative domains from the domain sequence families, 49% are in the range of 51-150 residues, whereas 64% of the representative chains over 200 residues have more than 1 domain. Of the representative chains, 8.5% are part of multichain domains. The largest multichain domain in the database has 14 chains and 1,400 residues, whereas the largest single-chain domain has 907 residues. The largest number of domains found in a protein is 13. The analysis shows that over the history of the PDB, new domain folds have been discovered at a slower rate than by random selection of all known folds. Between 1992 and 1997, a constant 1 in 11 new domains deposited in the PDB has shown no sequence similarity to a previously known domain sequence family, and only 1 in 15 new domain structures has had a fold that has not been seen previously. A comparison of the September 1996 release of 3Dee to the Structural Classification of Proteins (SCOP) showed that the domain definitions agreed for 80% of the representative protein chains. However, 3Dee provided explicit domain boundaries for more proteins. 3Dee is accessible on the World Wide Web at http://barton.ebi.ac.uk/servers/3Dee.html.

3Dee: a database of protein structural domains

The 3Dee database is a repository of protein structural domains. It stores alternative domain definitions for the same protein, organises domains into sequence and structural hierarchies, contains non-redundant set(s) of sequences and structures, multiple structure alignments for families of domains, and allows previous versions of the database to be regenerated.

AVAILABILITY

3Dee is accessible on the World Wide Web at the URL http://barton.ebi.ac.uk/servers/3Dee.html.

Analytical description of backscattered states of polarization in polarization optical time-domain reflectometry measurements on uniformly twisted linearly birefringent optical fiber

We present a detailed derivation of the locus of Rayleigh backscattered states of polarization for polarization optical time domain reflectometry in uniformly twisted optical fiber with intrinsic linear birefringence. The locus is algebraically a quartic whose topology is determined by the relative orientation on the Poincaré sphere of the input SOP and the effective birefringence vector. We present an analysis that indicates how experimental data may be interpreted, through geometric parameters of the locus, for evaluating the fiber parameters. The analysis also indicates the minimum number of experimental data points required for meaningful values for the fiber parameters to be obtained.

JPred: a consensus secondary structure prediction server

An interactive protein secondary structure prediction Internet server is presented. The server allows a single sequence or multiple alignment to be submitted, and returns predictions from six secondary structure prediction algorithms that exploit evolutionary information from multiple sequences. A consensus prediction is also returned which improves the average Q3 accuracy of prediction by 1% to 72.9%. The server simplifies the use of current prediction algorithms and allows conservation patterns important to structure and function to be identified.

AVAILABILITY

http://barton.ebi.ac.uk/servers/jpred.h tml

CONTACT

geoff@ebi.ac.uk

Continuous and discontinuous domains: an algorithm for the automatic generation of reliable protein domain definitions

An algorithm is presented for the fast and accurate definition of protein structural domains from coordinate data without prior knowledge of the number or type of domains. The algorithm explicitly locates domains that comprise one or two continuous segments of protein chain. Domains that include more than two segments are also located. The algorithm was applied to a nonredundant database of 230 protein structures and the results compared to domain definitions obtained from the literature, or by inspection of the coordinates on molecular graphics. For 70% of the proteins, the derived domains agree with the reference definitions, 18% show minor differences and only 12% (28 proteins) show very different definitions. Three screens were applied to identify the derived domains least likely to agree with the subjective definition set. These screens revealed a set of 173 proteins, 97% of which agree well with the subjective definitions. The algorithm represents a practical domain identification tool that can be run routinely on the entire structural database. Adjustment of parameters also allows smaller compact units to be identified in proteins.

Covert hypothyroidism with weight loss and atrial fibrillation

Hypothyroidism may present with weight gain and/or cardiovascular manifestations such as bradycardia or cardiac failure, but has not previously been documented as presenting with atrial fibrillation and weight loss. Our case highlights the importance of thyroid function tests in heart failure and emphasises the importance of regular follow-up after irradiation to the thyroid.

The catabolism and heterotrophic nitrification of the siderophore deferrioxamine B

Three bacteria, two of which were previously noted as active heterotrophic nitrifiers, were examined for their ability to grow and nitrify with the siderophore deferrioxamine B as the carbon source. Pseudomonas aureofaciens displayed limited growth and nitrification while a heterotrophic nitrifying Alcaligenes sp. was without action concerning its metabolism of deferrioxamine B. The third bacterium, a unique Gram-negative soil isolate, was unable to nitrify deferrioxamine B but grew well when the siderophore was employed as the sole C source. The Gram-negative bacterium removed deferrioxamine B from the medium and left only residual amounts of the compound in solution at the termination of its growth. The organism was without action when the ferrated analogue of deferrioxamine B, ferrioxamine B, served as either the C source for growth, for metabolism by resting cells, or as the substrate for cell-free extracts. Deferrioxamine B, by contrast, was rapidly metabolized by resting cells. Cell-free extracts of the bacterium synthesized a monohydroxamate(s) when deferrioxamine B was the substrate. The catabolism of deferrioxamine B, which is synthesized by soil microbes, suggests that soil microflora have the ability to return deferrioxamine B, and perhaps other, siderophores to the C cycle.