Posts Tagged ‘gpmtg’

Protein-structure-guided discovery of functional mutations across 19 cancer types : Nature Genetics : Nature Research

December 11, 2016

Protein-structure-guided discovery of functional mutations across 19 #cancer types http://www.nature.com/ng/journal/v48/n8/abs/ng.3586.html Cancer3D relates SNVs to drugs

http://www.nature.com/ng/journal/v48/n8/abs/ng.3586.html

Skhizein: a really nice animation on today’s grpmtg

December 9, 2016

https://vimeo.com/36824575

PLOS Computational Biology: PredictSNP: Robust and Accurate Consensus Classifier for Prediction of Disease-Related Mutations

November 28, 2016

PredictSNP…Consensus Classifier for Prediction of Disease-Related
Mutations http://journals.PLOS.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003440 Demo of various #ensemble approaches

Frustration in biomolecules | Quarterly Reviews of Biophysics | Cambridge Core

November 1, 2016

#Frustration in biomolecules
https://www.cambridge.org/core/journals/quarterly-reviews-of-biophysics/article/frustration-in-biomolecules/DECEA176849986FC11DB079C1EB4B24A Reviews the field: how large molecules pay a local price to achieve global stability

PLOS Computational Biology: PredictSNP2: A Unified Platform for Accurately Evaluating SNP Effects by Exploiting the Different Characteristics of Variants in Distinct Genomic Regions

October 9, 2016

PredictSNP2: A Unified Platform http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1004962 Ensembles many scores for the impact of non-coding variants, including #FunSeq

Elucidating Molecular Motion through Structural and Dynamic Filters of Energy-Minimized Conformer Ensembles

September 30, 2016

Elucidating Molecular Motion (compatible w. #NMR relaxation times) through…Filters of Energy-Minimized…Ensembles
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3983377

Predicting peptide binding sites on protein surfaces by clustering chemical interactions – Yan – 2014 – Journal of Computational Chemistry – Wiley Online Library

July 4, 2016

Predicting peptide binding sites on protein surfaces by ACCLUSTER http://onlinelibrary.wiley.com/doi/10.1002/jcc.23771/abstract #Chemical interactions out perform pure #packing

Learning the Sequence Determinants of Alternative Splicing from Millions of Random Sequences: Cell

April 24, 2016

Learning the…Determinants of Alternative #Splicing [in a largely linear model] from Millions of Random Sequences
http://www.cell.com/cell/abstract/S0092-8674(15)01271-4

** Rosenberg et al Cell. 2015

Builds a model of splicing using a library of randomized sequence Also, builds a generalized model for predicting effect of a SNP in the Geuvadis RNAseq
7mer model does well with lots of data

Protein folds recognized by an intelligent predictor based-on evolutionary and structural information – Cheung – 2015 – Journal of Computational Chemistry – Wiley Online Library

April 17, 2016

Fold [class] recognized by an…[NN] predictor based-on evolutionary & structural info., w/ particle-swarm training
http://onlinelibrary.wiley.com/doi/10.1002/jcc.24232/full

Ngaam J. Cheung,
Xue-Ming Ding,
Hong-Bin Shen

First published: 27 October 2015
DOI: 10.1002/jcc.24232

PLOS Genetics: A Simple Model-Based Approach to Inferring and Visualizing Cancer Mutation Signatures

February 27, 2016

Model-Based Approach to Inferring…#Cancer Mutation Signatures http://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1005657 Assuming independence betw 3 NTs, 11 v 95 parameters

QT:{{”
The first contribution of this paper is to suggest a more parsimonious approach to modelling mutation signatures, with the benefit of producing both more stable estimates and more easily interpretable signatures. In brief, we substantially reduce the number of parameters per signature by breaking each mutation pattern into “features”, and assuming independence across mutation features. For example, consider the case where a mutation pattern is defined by the substitution and its two flanking bases. We break this into three features
(substitution, 3′ base, 5′ base), and characterize each mutation signature by a probability distribution for each feature (which, by our independence assumption, are multiplied together to define a distribution on mutation patterns). Since the number of possible values for each feature is 6, 4, and 4 respectively this requires 5 + 3 + 3 = 11 parameters instead of 96 − 1 = 95 parameters. Furthermore, extending this model to account for ±n neighboring bases requires only 5 + 6nparameters instead of 6 × 42n − 1. For example, considering ±2 positions requires 17 parameters instead of 1,535. Finally,
incorporating transcription strand as an additional feature adds just one parameter, instead of doubling the number of parameters. “}}