[tags scilit,jclub,GG,fromjclub,conferences,psb18]

Cell-specific prediction and application of drug-induced gene expression Extracting a biologically relevant latent space from cancer transcriptomes with variational autoencodersFunctional network community detection can disaggregate and filter multiple underlying pathways in enrichment analysesA heuristic method for simulating open-data of arbitrary complexity that can be used to compare and evaluate machine learning methodsCoalitional game theory as a promising approach to identify candidate autism genesNetwork analysis of pseudogene-gene relationships: from pseudogene evolution to their functional potentialsAn ultra-fast and scalable quantification pipeline for transposable elements from next generation sequencing data

Mount Sinai is organizing a symposium in Pamela’s memory that will take place on April 16. Please see attached flyer for more information

###############################

Quantum Computing

###############################

QC call minutes 1/11/18:

· Somatic Mosaicism in Brain project: consequences for SCZ (others in Consortium for ASD, Tourette’s and Epilepsy); specifically, CNVs in single cells; 300 cells per individual, 5 SCZ, 5 CTRLs, 4000 data points per cell, seeking SVM for data classification, or significant features such as genes relevant to SCZ; reduce the resolution and coarse-grain to meet current size limitations of IBM Qubit system – 50 qubits.

· Kanix Wang, U of Chicago (Andrey Rezesky’s lab) Classification of common human determinants: separate genotypic influence from environment and combine to produce phenotypes; define “Disease Liability” that accounts for the genotypic influence (Gaussian distributed); correlation between disease pairs; issues with large sample size (data on 129K families), reduced through filtering, as well as large numbers of disease, reduced as well; model involves decomposition of variance using genetic + several environmental factors, variance among couples, siblings and families; determine heritability; lower size to meet limitations

· Mike Gandal (UCLA) integrative models of psychiatric disease genomic regulation: integration of genotypes, RNA-seq, QTLs into single network and connect this with phenotype; current approach WGCNA; QC could help with non-linearity aspect, small-size + high complexity is sort of problem to deal with;

· Dimension of available space 4^(N-1), N = number of qubits, so lots of problem space, but there are other bottlenecks to large-scale tasks: T1 energy leak (weaker) < quantum coherence < measurement fidelity (environmental noise, precision of acquisition) < gate fidelity (bad control over phase) are issues, need better control; max 100-200 gates, this limits complexity of feature space; cannot debug as this causes wavefunction collapse

· Gerstein lab to come up with proposal presentation for next call