Long time reader, first time poster.
Quick question for those you you our there with some more long-term practical experience.
TLDR: Large amount of unmapped reads map as salmon. Can using salmon sperm DNA to preclear the ChIP beads be getting into the final sequenced sample?
I have a ChIP-seq experiment where we have input and 3 targets. In 2 of the samples (both from the same target) we have a very low mapping rate (Bowtie2: 50-30% mapping). I extracted the unmapped reads (samtools f 4) and blasted a couple of them to get an idea of what it was and i got two different species of salmonoids. I got the FASTA for one of them and ran fastq-screen and about 50% of the reads match. I suspect (but am still working on it) that the rest of the unmapped sequences come from another salmonoid. I, of course, am suspicious that this is from the salmon DNA we use to preclear the beads (because where else??). All of our beads for all of our samples are precleared like this though and only 2 samples have this problem, but my thought is that perhaps if the ChIP was not so successful perhaps the salmon DNA is enough to out compete the sample DNA. All the people in the wetlab side of things are positive this cannot be the case... Is this a possibility? Anyone else ever have this problem?
To be sure take all unmapped reads and map against a salmon reference genome to get a percentage. If this is notably high then it essentially does not matter what other people think, then it is likely that the blocking DNA contaminated the libraries. Is this a good antibody, and are the peaks decent? The poorer the IP the more prevalent any contamination or unspecificly pulled-down DNA will be in the library. ChIP is a pain, I know by experience. On the other hand if you simply discard the unmapped reads and the quality of the remaining library is ok then things might still be usable. It is not uncommon to use carrier DNA during library prep if working with low inputs, e.g. DNA from E. coli to have a stable amount of DNA to start the prep with. Eventually you then have to sequence deeper as a fraction of librar will inevetably by just carrier DNA.