Is it normal to hav very low mapping for TREATMENT file in Chip-seq?

[anibhax]

Hi,

I'm working on chip-seq and it looks like I have an issue. I'm not sure if it's normal for a chip-seq analysis.

I have a control and treatment file (in fastq format), which I mapped it using Bowtie2. The problem here is, while the input file has 90% mapped, treatment has only 10% mapped (mapping statistics by samtools flagstat option). Is it something I should be worried about?

[kopi-o]

Yes, it sounds way too low.

[anibhax]

So normally there should not be any significant changes in mapping statistics between control and treatment right?

[zinky]

make sure that u have done the quality controll step ,this pheno may result from adaptor contamination

[kopi-o]

So normally there should not be any significant changes in mapping statistics between control and treatment right?

If the ChIP has succeeded, typically no. But I think it's not uncommon to get poor enrichment leading to things like adapter contamination as mentioned above. Or perhaps the sequencing was suboptimal so you have poor quality at the end of the reads and Bowtie has trouble mapping with certain parameters.

[anibhax]

make sure that u have done the quality controll step ,this pheno may result from adaptor contamination

@Zinky: I already did the quality check. I don't think it is adaptor contamination!

[anibhax]

If the ChIP has succeeded, typically no. But I think it's not uncommon to get poor enrichment leading to things like adapter contamination as mentioned above. Or perhaps the sequencing was suboptimal so you have poor quality at the end of the reads and Bowtie has trouble mapping with certain parameters.

When I ran the FastQC analysis, I checked all those adaptor contamination. I used trim-galore too to remove poor quality reads and default Illumina adaptors. Still, it gives me poor results. Hence, the bemusement.

I was wondering what might be the cause. The difference between treatment and control seem to be too much!

[Chipper]

Was the control IgG or input? If only 10% align I would suspect a contamination from the beads, perhaps they were blocked with ssDNA or you have osm bacteria growing in the buffer. Try a quick assembly using e.g Minia and blast the major contigs.

The 10% could still be usable if you have good enrichment.

[anibhax]

Was the control IgG or input? If only 10% align I would suspect a contamination from the beads, perhaps they were blocked with ssDNA or you have osm bacteria growing in the buffer. Try a quick assembly using e.g Minia and blast the major contigs.

The 10% could still be usable if you have good enrichment.

The control was Input. Does it make a difference in the analysis if it is input or IgG?

[GenoMax]

As chipper suggested in #8, have you tested to eliminate the possibility that the non-mapping fraction is some kind of contaminant?

This software package from folks at Babraham can help: http://www.bioinformatics.babraham.a.../fastq_screen/

[Chipper]

The control was Input. Does it make a difference in the analysis if it is input or IgG?

You don't use beads for input. If they are the source of the contaminant you would expect to see it in IgG but not in input.

[anibhax]

You don't use beads for input. If they are the source of the contaminant you would expect to see it in IgG but not in input.

I see. Well, I did BLAST the reads (a handful) which were unmapped. They mapped to human with a significant e-value. But I'm working on a worm. I'm not sure what to make out of the BLAST results.

[anibhax]

As chipper suggested in #8, have you tested to eliminate the possibility that the non-mapping fraction is some kind of contaminant?

This software package from folks at Babraham can help: http://www.bioinformatics.babraham.a.../fastq_screen/

Hi GenoMax,

For the fastq_screen I'm guessing it's necessary to know which adaptors or vectors I think might be involved prior to using it. Is it correct?

If so, are there default adapters/vectors available for certain Illumina platforms or do I need to contact the people who ran the analysis for the list?