#!/bin/bash

DATA=.
SCRIPTS=../../../scripts

# Change this to the directory where Circos is installed 
CIRCOS=/home/martink/work/circos/svn/

# grep treats special characters as normal unless they're escaped
#
#https://stackoverflow.com/questions/6775904/grepping-using-the-alternative-operator

# About 43,000 genes that are on canonical chromosomes
echo "Creating list of 43,000 genes"
cat $DATA/genes.human.refseq.txt | grep -v "#" | grep -v -i "random\|_alt\|none\|chrun" > genes.43000.txt

# Gene entries with unique names
echo "Creating list of 19,300 genes (unique names)"
cat genes.43000.txt | sort -u -k 13,13 > genes.19300.txt

# Prefixes
echo "Counting genes for each 3-letter prefix"
cat track.genes.txt | cut -d " " -f 5 | sed 's/name=//' | cut -c 1-3 | sort | uniq -c | sort -rn > prefixes.count.txt

# Gene entries with unique first 3 letter substrings in their name
echo "Creating list of 3,500 genes (unique 3-letter prefix)"
cat genes.19300.txt | awk '{print $13,$0}' | uniq -w 3 > genes.3500.txt

# Histogram of gene size
echo "Creating histogram of gene sizes (in kb)"
cat genes.19300.txt | awk '{print ($6-$5+1)/1000}' | $SCRIPTS/histogram -min 0 -max 200 -binsize 10 > histogram.gene.size.txt

# Number of genes on each chromosome
echo "Counting genes on each chromosome"
cat genes.19300.txt | cut -d $'\t' -f 3 | sort | uniq -c | sort -nr > count.bychr.txt

# Create a gene track with fields
# CHR START END NUM_EXONS name=GENENAME
echo "Creating Circos data file of gene positions"
cat genes.19300.txt | awk '{print $3,$5,$6,$9,"name="$13}' | sed 's/chr/hs/' > track.genes.txt

# Count 
BINS="1 5 10 20"
for ws in $BINS; do
  echo "Creating Circos data file of gene counts in $ws Mb windows"
  cat track.genes.txt | $CIRCOS/tools/resample/bin/resample -bin ${ws}e6 -count > track.gene.count.${ws}mb.txt
done