MT @Rbaltman Statins by Numbers http://nyti.ms/1bwpXzc Has medical #education prepared most doctors for the discipline’s money-ball moment?
http://www.nytimes.com/2013/11/30/opinion/statins-by-numbers.html
Summary for:
“Epigenomic Alterations in Localized and Advanced Prostate Cancer” Lin PC, Giannopoulou E, Park K, Mosquera JM, Sboner A, Tewari AK, Garraway LA, Beltran H, Rubin MA*, Elemento O*. 2013. Epigenomic alterations in localized and advanced prostate cancer. Neoplasia
http://www.ncbi.nlm.nih.gov/pubmed/23555183
In this paper, the authors take advantage of new advances in reduced representation bisulfite sequencing, a method for measuring DNA methylation patterns genome-wide, with high coverage and
single-nucleotide resolution, to study methylation patterns in prostate cancer. Working with a prostate cancer cohort already studied with DNA-Seq and RNA-Seq analyses, the authors identified
differentially methylated regions (DMRs), comparing the methylation of prostate cancer samples to benign prostate samples. The analysis found an increase in DNA methylation in prostate cancer samples, and that the methylation was more diverse and heterogeneous compared to the patterns of benign samples. Furthermore, it was found that genes near hypermethylated DMRs tended to have decreased expression, while genes near hypomethylated DMRs tended to have increased expression. Additional analyses revealed that breakpoints associated with prostate-cancer-specific deletions, duplications, and translocations tended to be highly methylated in benign prostate tissue. Finally, a study of CpG islands at different stages of prostate cancer (benign vs. PCa vs. CRPC (castration-resistant prostate cancer)) revealed that certain islands become increasingly methylated with disease severity. The authors used this data as the basis for two classification models: one to discriminate between benign prostate tissue and PCa tissue, and another to discriminate between PCa tissue and CRPC tissue. Both models demonstrated high sensitivity and specificity, indicating that CpG islands with high discriminatory power could serve as a diagnostic basis for predicting disease aggressiveness. Finally, additional analyses revealed that breakpoints associated with
prostate-cancer-specific deletions, duplications, and translocations tended to be highly methylated in benign prostate tissue.
http://nar.oxfordjournals.org/content/40/D1/D930.long
HaploReg explores functional annotations, such as chromatin states in varied cell types, sequence conservation, regulatory motif
alterations and eQTLs, of linked SNPs or indels within LD block of queried SNPs. The output provides a the guide to develop hypotheses of functional impact of non-coding variants, especially GWAS SNPs. HaploReg is currently limited to known variants (e.g. 1000 Genome variants and dbSNPs) and is unable to deal with private variants.
Can the Climate Corporation Help Farmers Survive? Mining #bigdata changes a most traditional profession
http://www.newyorker.com/reporting/2013/11/11/131111fa_fact_specter MT @PernilleT
ANNALS OF SCIENCE
CLIMATE BY NUMBERS
Can a tech firm help farmers survive global warming?
BY MICHAEL SPECTER
NOVEMBER 11, 2013
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).
#CEO, COO,CFO,CIO,CTO,CSO, #CDO….Ugh! Ogilvy Chief Data Officer Role May Be Sign of Things to Come
http://adage.com/article/datadriven-marketing/ogilvy-chief-data-officer-role-sign-things/243713 MT @KirkDBorne
http://adage.com/article/datadriven-marketing/ogilvy-chief-data-officer-role-sign-things/243713/
What College Rankings Really Tell Us : The New Yorker
THE ORDER OF THINGS
BY MALCOLM GLADWELL
http://www.newyorker.com/reporting/2011/02/14/110214fa_fact_gladwell
This is a very interesting article that illustrates the importance of considering the weighting various of features. It gives some very concrete examples of how one can get very different rankings by picking weights. All of this is in a sense obvious but it is nice to see spelled out with regard to school rankings.
Simpson’s Paradox That Both Raises and Reduces Wages
http://lifehacker.com/four-common-statistical-misconceptions-you-should-avoid-906056582
MT @KromerBigData: #DataScience in @nytimes http://bit.ly/13mPhxS McKinsey says ~60% more in field needed for ~500K new jobs in 5yrs
QT:”As an academic field, it cuts across disciplines, with courses in statistics, analytics, computer science and math, coupled with the specialty a student wants to analyze, from patterns in marine life to historical texts.”
QT:”Universities can hardly turn out data scientists fast enough. To meet demand from employers, the United States will need to increase the number of graduates with skills handling large amounts of data by as much as 60 percent, according to a report by McKinsey Global Institute. There will be almost half a million jobs in five years, and a shortage of up to 190,000 qualified data scientists, plus a need for 1.5 million executives and support staff who have an understanding of data.”
Looks like it click right into out database and do some powerful visualizations… assume Excel can do most of this as well.
http://www.visokio.com/