Abstract
Hi-C and chromatin immunoprecipitation (ChIP) have been combined to identify long-range chromatin interactions genome-wide at reduced cost and enhanced resolution, but extracting information from the resulting datasets has been challenging. Here we describe a computational method, MAPS, Model-based Analysis of PLAC-seq and HiChIP, to process the data from such experiments and identify long-range chromatin interactions. MAPS adopts a zero-truncated Poisson regression framework to explicitly remove systematic biases in the PLAC-seq and HiChIP datasets, and then uses the normalized chromatin contact frequencies to identify significant chromatin interactions anchored at genomic regions bound by the protein of interest. MAPS shows superior performance over existing software tools in the analysis of chromatin interactions from multiple PLAC-seq and HiChIP datasets centered on different transcriptional factors and histone marks. MAPS is freely available at https://github.com/ijuric/MAPS.
Highlights
While millions of candidate enhancers have been predicted in the human genome, annotation of their target genes remains challenging, because enhancers do not always regulate the closest gene in the linear genome sequence [1]
PLAC-seq/HiChIP datasets suffer from the biases introduced by differential effective fragment length, GC content and sequence uniqueness that are common to all 3C based methods [11], and contain the biases introduced during the chromatin immunoprecipitation (ChIP) procedure (i.e., ChIP enrichment level)
Mango considers the ChIP-introduced biases, application of Mango to PLAC-seq/ HiChIP data is still problematic for two reasons: 1) PLAC-seq/HiChIP enables detection of valid chromatin interactions between protein-bound regions and non-binding regions, which are not considered by Mango; 2) even for the detection of long-range interactions between two protein-bound regions, Mango is suboptimal since the anchor regions defined by MACS2 suffer from high false positive rate due to PLAC-seq/HiChIP-specific bias
Summary
While millions of candidate enhancers have been predicted in the human genome, annotation of their target genes remains challenging, because enhancers do not always regulate the closest gene in the linear genome sequence [1]. PLAC-seq [5] and HiChIP [6] technologies combine in situ Hi-C and chromatin immunoprecipitation (ChIP) to efficiently capture chromatin interactions anchored at genomic regions bound by specific proteins or histone modifications, achieving Kb resolution with fewer sequencing reads and much reduced sequencing cost [7] (Note 1 in S1 Text). Several software tools, including Fit-Hi-C [8], HiCCUPS [4], Mango [9] and hichipper [10] have been used to identify long-range chromatin interactions from PLAC-seq and HiChIP data. Mango considers the ChIP-introduced biases, application of Mango to PLAC-seq/ HiChIP data is still problematic for two reasons: 1) PLAC-seq/HiChIP enables detection of valid chromatin interactions between protein-bound regions and non-binding regions, which are not considered by Mango; 2) even for the detection of long-range interactions between two protein-bound regions, Mango is suboptimal since the anchor regions defined by MACS2 suffer from high false positive rate due to PLAC-seq/HiChIP-specific bias. Hichipper still relies on the statistical model in Mango to identify long-range interactions and is not designed to call interactions between protein binding regions and non-binding regions (more discussions in Note 2 in S1 Text)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
