feat: added note on UCSC liftover

docs/UCSC_liftover

@@ -0,0 +1,35 @@
+# UCSC liftover tool
+
+[UCSC liftover](https://genome.ucsc.edu/cgi-bin/hgLiftOver)
+
+## Issue: separate peaks map to same coordinates after liftover
+
+- When using the tool to liftover regions from hg19 to hg38, separate peaks in specific regions of hg19 map to overlapping coordinates after liftover.
+- This is because some contigs were not carried forward from hg19 to hg38, due to problems with the hg19 assembly that were resolved in hg38
+- All lifted over peaks with overlapping coordinates should be removed from analysis because the peaks were called on an uncertain region in the old genome assembly
+- Issues like this likely apply to all genome updates, not only human
+- The same issue also applies with [Ensembl assembly converter](https://www.ensembl.org/Homo_sapiens/Tools/AssemblyConverter) 
+
+## Solution
+Remove any peaks with overlapping coordinates after liftover before using the lifted over peak file:
+
+```bash
+#!/bin/bash
+
+module load bedtools
+
+EXTL=("../data/external/")
+
+# Sort the lifted over peakfile for use with bedtools
+sort -k1,1 -k2,2n ${EXTL}wong_fig3c_peaks_GRCh38.bed > ${EXTL}peaks.tmp && mv ${EXTL}peaks.tmp ${EXTL}wong_fig3c_peaks_GRCh38.bed
+
+# Bedtools merge count rows contributing to merged peaks (overlapping peaks will have count > 1)
+bedtools merge -i ${EXTL}wong_fig3c_peaks_GRCh38.bed -c 1 -o count > ${EXTL}counted.bed
+
+# Get non-overlapping peaks
+awk '/\t1$/{print}' ${EXTL}counted.bed > ${EXTL}filtered.bed
+
+# Intersect original file with non-overlapping peaks and output overlapping peaks
+bedtools intersect -wa -a ${EXTL}wong_fig3c_peaks_GRCh38.bed -b ${EXTL}filtered.bed > ${EXTL}wong_fig3c_peaks_GRCh38_correct_liftover.bed
+bedtools intersect -v -a ${EXTL}wong_fig3c_peaks_GRCh38.bed -b ${EXTL}filtered.bed > ${EXTL}wong_fig3c_peaks_GRCh38_overlapping.bed
+```