Does somebody know, if there are specialized databases for Hi-C raw data (paired-end reads)? Of course, there are many raw Hi-C datasets on GEO, but I failed to find any specialized Hi-C database on the Web.
I wouldn't expect anyone to duplicate storage of raw reads beyond what SRA and GEO do. But as far as aggregating multiple experiments, the WashU Epigenome Browser has long-range chromatin interaction experiment tracks available. Currently there are 73 tracks available for human, 27 each for mouse and fly.
We have been working on an initiative to host and visualize Hi-C data. We download the data from SRA and processes it with our pipeline to generate downloadable matrix files (stored on hdf5 files). So far we have some data for drosophila, mouse and human available.
There is a LOGIQA (Long-range genome interactions quality assessment) database. However it hosts only quality scores (assessed over Hi-C, 4C, ChIA-PET, Capture-C, etc.). Nevertheless it is very convenient for searching data you look for.
I don't believe there are any specialized Hi-C databases on the web...
(perhaps that will change with ENCODE3 + 4D Nucleome efforts)
Though Ryan is correct - I doubt anyone would duplicate what is already stored on GEO/SRA. Most paper submissions (GEO) include not only the fastq files but also some form of the interaction data. The data will either be in a tsv (3 column) format, tsv (matrix) format, or hdf5.
I think GEO will be your best bet for the time being.