E-GEOD-34073 - Uniform optimal framework for integrative next-gen sequence analysis

Released on 10 June 2013, last updated on 4 May 2014
Mus musculus
Samples (4)
Protocols (5)
Here, we have collapsed multiple analysis problems into two coherent categories, signal detection and signal estimation and adapted linear-optimal solutions from signal processing theory. Our algorithms for detection (DFilter) and estimation (EFilter) extend naturally to integration of multiple datasets. In benchmarking tests, DFilter outperformed assay-specific algorithms at identifying promoters from histone ChIP-seq, binding sites from transcription factor (TF) ChIP-seq and open chromatin regions from DNase- and FAIRE-seq data. EFilter similarly outperformed an existing method for predicting mRNA levels from histone ChIP-seq data (Spearman correlation: 0.81 - 0.89). We performed H3K4me3 and H3K36me3 ChIP-seq on e11.5 mouse forebrain and used DFilter and EFilter to predict promoters and developmental gene expression, uncovering plausible gene targets for SNPs associated with neurodevelopmental disorders. Generated two histone modifiction ChiP-seq in developing embryo mouse forebrain and using them for making bioligical inferences
Experiment type
Vibhor Kumar <kumarv1@gis.a-star.edu.sg>, MURATANI Masafumi, Shyam Prabhaker
Exp. designProtocolsVariablesProcessedSeq. reads
Investigation descriptionE-GEOD-34073.idf.txt
Sample and data relationshipE-GEOD-34073.sdrf.txt
Processed data (2)E-GEOD-34073.processed.1.zip, E-GEOD-34073.processed.2.zip