linkstream2 microblog

Genome Biol. 2012 Dec 13;13(12):R115.

Whole-genome reconstruction and mutational signatures in gastric cancer. Nagarajan N, Bertrand D, Hillmer AM, Zang ZJ, Yao F, Jacques PE, Teo AS, Cutcutache I, Zhang Z, Lee WH, Sia YY, Gao S, Ariyaratne PN, Ho A, Woo XY, Veeravali L, Ong CK, Deng N, Desai KV, Khor CC, Hibberd ML, Shahab A, Rao J, Wu M, Teh M, Zhu F, Chin SY, Pang B, So JB, Bourque G, Soong R, Sung WK, Tean Teh B, Rozen S, Ruan X, Yeoh KG, Tan PB, Ruan Y.

http://www.ncbi.nlm.nih.gov/pubmed/23237666

Some thoughts, much from WC:

Looks like the data is freely available via GEO ID : GSE30833 http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE30833

The article by Nagarajan et al. highlights the authors efforts to utilize de novo genome assembly of gastric cancer genomes to detect not only single nucleotide variants (SNV’s) and short
insertions/deletions (indels), but also larger scale genomic structural variation (SV) that could be signatures of cancer genomes. It is to be applauded that this is a whole genome analysis.

The authors present several interesting findings such as enrichment for C->A and T->A mutations in both cancer genomes, enrichment for C->A and C->T mutations in the H. pylori infected cancer genome (evidence of cytosine specific transcription mediated DNA repair due to deamination), and amplification and deletion of regions on chromosome 12 in the non-H. pylori infected genome.

Although copy number variants (CNV) could potentially be detected by exome sequencing alone, whole genome sequence enables the precise localization of such events, as well as the detection of variation in non-coding regions.

Their methodology relies on combining high-throughput short-read sequencing with longer DNA-PET (paired end tags) in order to construct higher confidence de novo assemblies with longer contiguous regions.

The opposite of clique completion: #Network deconvolution.. to distinguish direct dependencies http://go.nature.com/dVzNwC via @taziovanni

Network deconvolution as a general method to distinguish direct dependencies in networks

Soheil Feizi, Daniel Marbach, Muriel Médard & Manolis Kellis

http://www.nature.com/nbt/journal/vaop/ncurrent/full/nbt.2635.html

My thoughts:

Indirect relationships in a network can confound the inference of true direct relationships in a network. T, so this paper sought to develop a quantitative framework, termed network deconvolution (ND), to infer direct relationships and remove false positives in a network by quantifying and then removing indirect transitive relationship effects. The mathematical framework assumes that (1) an indirect relationship (edge) can be approximated as the product of its component direct edges and that (2) the observed edge weights are the sum of the direct and indirect edge weights – a linear dependency. The main application seems to be in mutual information (MI) and
correlation-based (COR) networks. They applied ND to various scenarios such as local network connectivity prediction (FFL
prediction), gene regulatory network prediction (in E. coli), prediction of interacting amino acids in protein structures (MI network) and coauthorship relationship network and found that (1) it can be used with various networks beyond just MI and COR (2) it can be used alone or more powerfully in combination with existing
methods/algorithms to improve predictions. In a sense it is the opposite of clique and module completion approaches (such as k-core).

This paper has: (1) Large-scale datasets compiled from literature and databases, (2) comprehensive gold standards for positive and negative samples, (3) a classifier algorithm (regularized Bayesian), and (4) further analysis beyond “functional prediction”, including an interaction network. It predicts a list of genes having some possible functions, and the authors have experimentally validated them.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2694471/

Genome Res. 2009 Jun;19(6):1093-106. Epub 2009 Feb 26.
Exploring the human genome with functional maps.
Huttenhower C, Haley EM, Hibbs MA, Dumeaux V, Barrett DR, Coller HA, Troyanskaya OG.

Large-scale genomic datasets can easily be transformed into various networks. The authors aimed to infer for each particular edge, whether or not it shows up in a particular tissue by training a model based on well curated tissue-specific expression as gold standards. The algorithm arrives at different tissue-specific networks from large-scale genomics datasets; without surprise, tissue-specific networks are more informative in predicting genes corresponding to diseases related to that particular tissue. For instance, a
testis-specific network performs better in predicting genes associated with male fertility phenotypes.

http://www.ploscompbiol.org/article/info%3Adoi%2F10.1371%2Fjournal.pcbi.1002694 PLoS Comput Biol. 2012 Sep;8(9):e1002694.
doi: 10.1371/journal.pcbi.1002694. Epub 2012 Sep 27.
Tissue-specific functional networks for prioritizing phenotype and disease genes.
Guan Y, Gorenshteyn D, Burmeister M, Wong AK, Schimenti JC, Handel MA, Bult CJ, Hibbs MA, Troyanskaya O

This paper introduces a new method for detecting copy number variants in cancer genomes that addresses deficiencies of previous detection methods. The new method, dubbed HHCRF by the authors, adds the use of sequential correlations in selecting classification features for inferring copy numbers and identifying clinically relevant genes. This improvement results in higher accuracy on noisy data, and the identification of more clinically relevant genes, relative to previous methods. These results were obtained by testing HHCRF on both simulated array-CGH microarray data, and on actual breast cancer, uveal melanoma, and bladder tumor datasets.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2677736/
Bioinformatics. 2009 May 15;25(10):1307-13. Epub 2008 Dec 3. Aneuploidy prediction and tumor classification with heterogeneous hidden conditional random fields.
Barutcuoglu Z, Airoldi EM, Dumeaux V, Schapire RE, Troyanskaya OG.

Summary adapted from from Koon-Kiu (KKY):

This paper studied cyclic motifs (cycles) in biological and
technological networks. A cycle can be characterized by the number of clockwise and counter-clockwise links, the number of pass-through nodes and the number of sources/sinks, etc. Direct counting of cycles of various length suggests a dependence between neighboring links, and such dependence is modeled by an interacting spin model. Fitting to the spin model shows that neighboring links tend to be in opposite directions (antiferromagnetic), resulting in a depletion of feedback loops in networks. Stability analysis concluded that the lack of feedback loop stabilizes the system in terms of perturbation around the fixed point.

Ma’ayan A, Cecchi GA, Wagner J, Rao AR, Iyengar R, Stolovitzky G. Ordered Cyclic Motifs Contributes to Dynamic Stability in Biological and Engineered Networks. Proceedings of the National Academy of Sciences 105, 19235-40 (2008) PMID: 19033453
http://ukpmc.ac.uk/abstract/MED/19033453/reload=0;jsessionid=aeku6lFJSKlT8wnr9czW.12

Summary adapted from Declan (DC):

The authors attempt to devise a few simple statistical metrics using high-throughput experimental data (many experiments involving immuno-precipitation coupled with mass spec) in order to predict protein-protein and domain-domain interactions involved in
transcription-related complexes. Each experiment entails using mass spec in order to identify the “prey” proteins that associate with a given “bait” protein. Broadly, protein-protein interactions between such prey proteins are predicted with statistical metrics that assign a likely interaction between a pair of proteins if that pair consistently co-occurs (in high abundance) across multiple
experiments. An example of one of their well-performing metrics is the Sorenson coefficient, which is the ratio of twice the number of experiments in which both proteins occur to the number of experiments in which either or both of these proteins occur (naively, this can be thought of as the degree of intersection between the experiments in which Protein A occurs and the experiments in which Protein B occurs). Using the top 10% of predicted interactions for each of their 4 statistical metrics, they validate many interactions with data from the literature, and they also perform experimental validation and docking studies in order to validate a tiny number of their
predictions. They supply their resultant networks as web-accessible data files.

Mazloom AR et al.
Recovering Protein-Protein and Domain-Domain Interactions from Aggregation of IP-MS Proteomics of Coregulator Complexes. PLoS Computational Biology. 7, e1002319 (2011) PMID: 22219718
http://www.ploscompbiol.org/article/info%3Adoi%2F10.1371%2Fjournal.pcbi.1002319

Functions of Bifans in Context of Multiple Regulatory Motifs in Signaling Networks
Azi Lipshtat, , Sudarshan P. Purushothaman, Ravi Iyengar and Avi Ma’ayan http://www.cell.com/biophysj/abstract/S0006-3495(08)70511-3
Biophys J. 94, 2566-2579 (2008) PMID: 18178648

Summary adapted from Chao (CC):

The authors constructed ordinary differential equations to model the quantitative dynamical behavior in an example bifan motif, in which two kinases p38alpha and JNK1 cross-regulate two transcription factors ATF2 and Elk-1. The simulation indicates that the bifan motif provide temporal regulation of signal propagation and can act as signal sorters, filters, and synchronizers. Bifan motifs with OR gate configurations mediate rapid responses, whereas the one with AND gate configurations introduces delays and allows prolongation of signal outputs. The authors also conducted several sets of simulations, using different initial conditions or considering bifan in a more complex network, and found that synchronization is a robust property of bifan motifs. This study makes a thorough investigation into the dynamical characteristics of the bifan motif based on decent mathematical models; however, there is no experimental result to further support those simulation results.