("Tenebre", iPhone6s, Camera.)
Simply being a part of traveling alone from the start, and somehow cohere.
□ 種分類と音楽様式における”classification”の概念と存在意義の儚さは似通っている。ヒュームの言葉を借りるまでもなく、因果が転写されたのではなく、同じ様式が群として隣接する。時間の矢ではなく、空間的に凝集することに本質があるのなら、認知は決定論的にプリントアウトされる。
□ わたしたちの心は、現し身に起きている事象とは全く違う所から来て、目にするものは単なる偶然の影遊びなのかもしれない。そして別の選択肢は辿られなかったのではなく、忘れているだけ。
□ 描かれざる被写体の背後、雲の彼方に星が群れている。輪郭があろうとなかろうと、私たちが目にしているものは同じ。額縁の内側も外側も等しく、凝集体で象られている。
□ ethereumが暴騰してる。ヘッジファンドと中国が、Brexitに備えて仮想通貨を買い集めているらしい。OpenDataとEthereumは相補性のある概念に基づいている。国家や組織以上に有用な資金分配のガバナンスを実現する共通の道筋を示す点において。
□ core: improved chain db performance by using sequential keys:
>> https://github.com/ethereum/go-ethereum/pull/2455
□ the DANN™ Distributed Autonomous Neural Network, a perceptron in ethereum
>> https://github.com/we-are-prot/blockbrain/blob/master/Perceptron.sol
□ SmidgION. Same @nanopore technology, docks w portable devices #nanoporeconf
>> https://www.nanoporetech.com/products-services/smidgion
Project Zumbador is also early development, a combined sample and library prep device that would deliver DNA captured on beads directly into the flow cell.
□ DRIMSeq: a Dirichlet-multinomial framework for multivariate count outcomes in genomics:
>> http://f1000research.com/articles/5-1356/v1
The Dirichlet-multinomial model naturally accounts for the differential gene expression w/o losing information about overall gene abundance and by joint modeling of isoform expression, it has the capability to account for their correlated nature. DRIMSeq has strong performance on transcript-level quantification w/ kallisto & even outperforms DEXSeq when the sample size is very small.
□ DeepCpG: Accurate prediction of single-cell DNA methylation states using deep learning:
>> http://biorxiv.org/content/biorxiv/early/2016/05/27/055715.full.pdf
additional data obtained from parallel methylation- transcriptome sequencing protocols to annotate regions w/ increased epigenetic diversity. integrate multiple datasets from parallel-profiling methods, which are now becoming increasingly available for different molecular layers.
□ Visualization and analysis of single-cell RNA-seq data by kernel-based similarity learning:
>> http://biorxiv.org/content/early/2016/05/09/052225
SIMLR demonstrates high sensitivity and accuracy on high-throughput PBMC data sets by the GemCode single-cell technology from 10x Genomics. t-SNE computes the similarity of the high-dimensional data points using a Gaussian kernel as a distance measure and projects the data onto a lower dimension that preserves this similarity.
□ Parallel computation of genome-scale RNA secondary structure to detect structural constraints on human genome
>> https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-016-1067-9
ParasoR directly computes the ratios of DP variables to avoid the reduction of numerical precision, caused by the cancellation of a large number of Boltzmann factors.
□ OPERA-LG: efficient and exact scaffolding of large, repeat-rich eukaryotic genomes with performance guarantees:
>> http://genomebiology.biomedcentral.com/articles/10.1186/s13059-016-0951-y
OPERA-LG has been exploited in several projects, incl the assembly of CHO cell lines using SOLiD mate-pair and Illumina paired-end datasets. The use of PacBio reads to aid scaffolding has also been demonstrated by custom methods designed for this task (SSPACE-LongRead and LINKS). For the large synthetic datasets, the GAGE pipeline was used to evaluate the final assemblies for contig errors and scaffold errors. In order to simultaneously use data from multiple “jumping” libraries for scaffolding, OPERA-LG uses a three-staged process, To reduce runtime, OPERA-LG pre-computes sum of contig lengths g, Q3+3×IQR at 500-bp intervals and returns the value of InlineEquation E(Sˆ)
□ Evaluation of hybrid and non-hybrid methods for de novo assembly of nanopore reads:
>> http://bioinformatics.oxfordjournals.org/content/early/2016/05/08/bioinformatics.btw237.abstract
NanoMark is a system for benchmarking DNA assembly tools, based on 3rd generation sequencers. benchmarked five non-hybrid (in terms of both error correction and scaffolding) assembly pipelines as well as two hybrid assemblers which use third generation sequencing data to scaffold Illumina assemblies. Results show that hybrid methods are highly dependent on the quality of NGS data, but much less on the quality and coverage of nanopore data and perform relatively well on lower nanopore coverages.
□ Spatial mnemonic encoding: Theta power decreases co-occur with medial temporal lobe BOLD:
>> http://biorxiv.org/content/early/2016/05/11/052720.full.pdf+html
Participants were trained to use two mnemonic encoding strategies: the spatial Method of Loci and the non-spatial pegword method. In both methods participants have to link internal cues, which are either familiar way points or semantic associations to numbers, to items presented during the encoding phase.
□ KimLabIDV: Application for Interactive RNA-Seq Data Analysis and Visualization:
>> http://biorxiv.org/content/biorxiv/early/2016/05/13/053348.full.pdf
For DE analysis, IDV implements a GUI for the Mann-Whitney U test and several popular DE pack- ages such as DESeq and SCDE. the cell-to-cell distances computed in the SCDE module can be used as distance measures for hierarchical clustering or tree generation. after community detection has been performed, users can conven- iently color samples by community in modules such as heatmap, PCA and t-SNE.
□ Hydrogen: Run code and get results inline using Jupyter kernels like IPython, IJulia, and iTorch
>> https://atom.io/packages/hydrogen
run your code directly in Atom using any Jupyter kernels.
□ TSCAN: Pseudo-time reconstruction and evaluation in single-cell RNA-seq analysis:
>> http://nar.oxfordjournals.org/content/early/2016/05/13/nar.gkw430.long
TSCAN uses a cluster-based minimum spanning tree (MST) approach to order cells. to evaluate TSCAN and other unsupervised pseudo-time reconstruction methods, using two time course data sets with multiple time points, HSMM and LPS, and intentionally avoided using any information on data collection time in our pseudo-time analyses.
□ VERSE: a versatile and efficient RNA-Seq read counting tool:
>> http://biorxiv.org/content/biorxiv/early/2016/05/14/053306.full.pdf
VERSE is more than 30x faster than HTSeq when computing the same gene counts. VERSE also supports a hierarchical assignment scheme, which allows reads to be assigned uniquely and sequentially to different types of features according to user-defined priorities.
□ Shedding light on the 'dark matter' of the genome
>> http://phys.org/news/2016-05-dark-genome.html
□ Global Mapping of Human RNA-RNA Interactions:
>> http://www.sciencedirect.com/science/article/pii/S109727651630106X
LIGR-seq: LIGation of interacting RNA followed by high-throughput sequencing. LIGR-seq data reveal unexpected interactions between small nucleolar (sno)RNAs and mRNAs, including those involving the orphan C/D box snoRNA, SNORD83B, that control steady-state levels of its target mRNAs.
□ Quantum field theory on curved spacetimes: Axiomatic framework and examples:
>> http://scitation.aip.org/docserver/fulltext/aip/journal/jmp/57/3/1.4939955.pdf
the point of view that the QFT model on a given spacetimes should be first constructed from “simple building blocks,” i.e., algebras associated to regions that are topologically simple and global structure is recovered from the properties of the net itself. The Chevalley-Eilenberg complex and the Koszul-Tate complex t together into one structure called the BV complex.
□ racon - Fast consensus module for raw de novo genome assembly of long uncorrected reads.
>> https://github.com/isovic/racon
□ Power considerations for λ inflation factor in meta-analyses of genome-wide association studies:
>> http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=10329187&fileId=S0016672316000069
□ Tellurium: A Python Based Modeling and Reproducibility Platform for Systems Biology:
>> http://biorxiv.org/content/biorxiv/early/2016/05/21/054601.full.pdf
Tellurium ensures exchangeability and reproducibility of computational models by supporting SBML (Systems Biology Markup Language), SED-ML (Simulation Experiment Description Markup Language), the COMBINE archive, and SBOL (Synthetic Biology Open Language). Tellurium also includes Antimony, a human-readable model definition language which can be converted to and from SBML.
□ Building a Kraken database with new FTP structure and no GI numbers:
>> http://www.opiniomics.org/building-a-kraken-database-with-new-ftp-structure-and-no-gi-numbers/
>> https://github.com/mw55309/Kraken_db_install_scripts/blob/master/download_human.pl
my $out = Bio::SeqIO->new(-file => ">human_genomic.tax.fna", -format => 'fasta');
my $newseq = Bio::PrimarySeq->new(-id => $id, -seq => $seq->seq, -desc => $seq->description);
□ SePIA: RNA and small RNA sequence processing, integration, and analysis:
>> http://biodatamining.biomedcentral.com/articles/10.1186/s13040-016-0099-z
SePIA has an optional pipeline, MirTPdb, to create, query & maintain an integrated SQLite database of predicted & validated miRNA-mRNA pairs. The consequent influx of available experimentally validated interactions could be easily included in the MirTPdb target database.
□ Efficient Integrative Multi-SNP Association Analysis via Deterministic Approximation of Posteriors.
>> http://www.ncbi.nlm.nih.gov/pubmed/27236919
“Deterministic Approximation of Posteriors” enables highly accurate joint enrichment analysis & identification of multiple causal variants. demonstrates by analyzing the cross-population eQTL data from the GEUVADIS project and the multi-tissue eQTL data from the GTEx project. genetic variants predicted to disrupt transcription factor binding sites are enriched in cis-eQTLs across all tissues.
□ JuPOETs: A Constrained Multiobjective Optimization Approach to Estimate Biochemical Model Ensembles in the Julia Programming Language
>> http://biorxiv.org/content/early/2016/05/30/056044
JuPOETs can be adapted to solve many problem types, including mixed binary and continuous variable types, bilevel optimization problems and constrained problems without altering the base algorithm.
□ The Go Ethereum Virtual Machine:
>> https://medium.com/@jeff.ethereum/go-ethereums-jit-evm-27ef88277520#.331k6qhhe
The JIT-EVM takes a different approach to running EVM byte-code and is by definition initially slower than the byte-code VM.
□ Data-driven hypothesis weighting increases detection power in genome-scale multiple testing:
>> http://www.nature.com/nmeth/journal/vaop/ncurrent/full/nmeth.3885.html
a method that assigns weights using covariates independent of the P-values under the null hypothesis but informative of each test's power or prior probability of the null hypothesis. find the optimal weight vector under a convex relaxation of the above optimization task, which in statistical terms corresponds to replacing the empirical cumulative distribution function of the p-values with Greenlander estimator. The resulting problem is convex and can be efficiently solved even for large numbers of hypotheses.
□ The SBOL Stack: A Platform for Storing, Publishing, and Sharing Synthetic Biology Designs:
>> http://pubs.acs.org/doi/abs/10.1021/acssynbio.5b00210
the inherent distributed querying functionality of RDF DBs can be used to allow different SBOL stack databases to be queried simultaneously
□ Differential analysis of RNA-Seq incorporating quantification uncertainty:
>> http://biorxiv.org/content/biorxiv/early/2016/06/10/058164.full.pdf
The method is implemented in an interactive shiny app called sleuth that utilizes kallisto quantifications and bootstraps.
□ Bayesian Markov models consistently outperform PWMs at predicting motifs in nucleotide sequences:
>> http://nar.oxfordjournals.org/content/early/2016/06/09/nar.gkw521.full
BaMMs achieve significantly (P=1/16) higher cross-validated partial AUC than PWMs in 97% of 446 ChIP-seq ENCODE & improve performance by 36%. zeroth-order BaMMs, which are simply PWMs trained with the standard EM-type algorithm as implemented in MEME and XXmotif. These were correlated with predictions from the deep learning method DeepBind and various other methods.
□ Navigating the phenotype frontier: The Monarch Initiative:
>> http://biorxiv.org/content/biorxiv/early/2016/06/15/059204.full.pdf
□ SourceData - a semantic platform for curating and searching figures:
>> http://biorxiv.org/content/biorxiv/early/2016/06/13/058529.full.pdf
□ Operant Behavior in Model Systems:
>> http://biorxiv.org/content/biorxiv/early/2016/06/13/058719.full.pdf
there is an inhibitory interaction between the world-learning and the self- learning processes during operant composite conditioning, such that only the effects of the world-learning process can be detected after operant composite conditioning.
computational neuroscientists unravel interactions between flexible model-based and efficient model-free learning processes. Animals and humans have evolved over millions of years to become experts in mastering this fundamental trade-off. Focusing on only one aspect of it fails to capture essential processes controlling behavior.
□ RiVIERA-MT: A Bayesian model to infer risk variants in related traits using summary statistics and functional genomic annotations
>> http://biorxiv.org/content/early/2016/06/16/059345
an efficient Markov Chain Monte Carlo approach by jointly sampling from the posterior distribution causal configurations for each locus and functional effects of each annotation that are shared among loci for the same trait and potentially correlate between traits.
□ Status of ELIXIR following UK referendum on EU
>> https://www.elixir-europe.org/news/status-elixir-following-uk-referendum-eu
EMBLもELIXIRも政府間の合意により設置された組織のため、EU離脱の直接的な影響は避けられるだろうとの見方。
□ MultiQC: A modular tool to aggregate results from bioinformatics analyses across many samples into a single report.
>> http://m.bioinformatics.oxfordjournals.org/content/early/2016/06/15/bioinformatics.btw354
>> http://multiqc.info
□ A direct multi-generational estimate of the human mutation rate from autozygous segments
>> http://biorxiv.org/content/biorxiv/early/2016/06/17/059436.full.pdf
frequent recurrence of mutations at polymorphic CpG sites, and an increase in C to T mutations in a 5’ CCG 3’ → 5’ CTG 3’ context in the Pakistani population compared to Europeans.
the count of singleton heterozygotes N0 to obtain the value 1.51 × 10^-8 ± 0.05 /bp/gen (= N0/LM). a single-nucleotide mutation rate of 1.41 ± 0.04 × 10^-8 /bp/gen and a non-crossover gene conversion rate of 8.75 ± 0.05 × 10^-6 /bp/gen.
□ Voodoo Machine Learning for Clinical Predictions:
>> http://biorxiv.org/content/biorxiv/early/2016/06/19/059774.full.pdf
evaluate the reliability of reported accuracies in studies that used machine learning & wearable sensor technology to clinical prediction. the accuracies reported by the studies using the wrong method (record-wise) were higher than the ones by correct (subject-wise) papers.
□ Streaming sequencing data from ENA for data-processing without wasting storage:
>> http://nxn.se/post/146066712225/streaming-rna-seq-data-from-ena