test PURECLIP peak calling

[slsevilla]

Determine difference in v1 and v2 of the pipeline using PURECLIP peak calling tool.

[slsevilla]

Completed testing with PureClip

Run info

General notes:

• no replicates
• dedup'ed bam is taken right after rule: dedup of the v2.0 iCLIP pipeline PureClip command:
• full code listed here: run_pureclip

  local pc_cmd="$pureclip_dir \
  -i $merged_bam \
  -bai ${merged_bam}.bai \
  -g $clean_reference \
  -ld -nt 8 \
  -o ${output_dir}/${sample_id}_${v_id}_sites.bed \
  -or ${output_dir}/${sample_id}_${v_id}_regions.bed"

Results

Ran with bedtools

bedtools jaccard -a $v1_sorted -b $v2_sorted > ${output_dir}/comparisons/${sample_id}_inboth.txt

Comparisons were as follows

• Control_Clip_notinV1.txt: 3807 (out of 3822)
• Control_Clip_notinV2.txt: 934 (out of 949)
• Ro_Clip_notinV1.txt: 3146 (out of 3178)
• Ro_Clip_notinV2.txt: 781 (out of 813)

Interpretation / Next steps

The results varied significantly from our current peakcalling method, with very few overlaps (<5%) overall. Given the satisfaction of the current peak calling method significant testing would need to be performed in order to justify this conversion. Seeing there has been no concern with current method, there isn't a justification to perform this change.

Closing this issue for now, and will potentially return to it at a later date.