{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "source bioinf_intro_config.sh\n",
    "rm -rf $IGV_DIR $STAR_OUT\n",
    "mkdir -p $TRIMMED $STAR_OUT $IGV_DIR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run with shorter intron limit"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "IGV needs indices for the BAM files.  The index allows it to quickly load reads from different parts of the genome."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for FASTQ in $RAW_FASTQS/27_MA_P_S38_L00[1]_R1_001.fastq.gz\n",
    "    do\n",
    "        FASTQ_BASE=\"$(basename ${FASTQ} '_001.fastq.gz')\"\n",
    "        echo \"---------------- TRIMMING: $FASTQ_BASE ----------------\"\n",
    "        fastq-mcf \\\n",
    "            $MYINFO/neb_e7600_adapters.fasta \\\n",
    "            $RAW_FASTQS/${FASTQ_BASE}_001.fastq.gz \\\n",
    "            -q 20 -x 0.5 \\\n",
    "            -o $TRIMMED/${FASTQ_BASE}_001.trim.fastq.gz\n",
    "        \n",
    "        echo \"---------------- MAPPING: $FASTQ_BASE ----------------\"\n",
    "        STAR \\\n",
    "            --runMode alignReads \\\n",
    "            --twopassMode None \\\n",
    "            --genomeDir $GENOME_DIR \\\n",
    "            --readFilesIn $TRIMMED/${FASTQ_BASE}_001.trim.fastq.gz \\\n",
    "            --readFilesCommand gunzip -c \\\n",
    "            --outFileNamePrefix ${STAR_OUT}/${FASTQ_BASE}_short_introns_ \\\n",
    "            --quantMode GeneCounts \\\n",
    "            --outSAMtype BAM SortedByCoordinate \\\n",
    "            --alignIntronMax 5000 \\\n",
    "            --outSJfilterIntronMaxVsReadN 500 1000 2000 \\\n",
    "            \n",
    "        echo \"---------------- INDEXING BAM: $FASTQ_BASE ----------------\"\n",
    "        samtools index ${STAR_OUT}/${FASTQ_BASE}_short_introns_Aligned.sortedByCoord.out.bam\n",
    "    done"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ls -ltr ${STAR_OUT}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Link Directory"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ln -s ${STAR_OUT}/*.bam* $GTF $FASTA $IGV_DIR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tarring"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tar --dereference \\\n",
    "    --create \\\n",
    "    --gzip \\\n",
    "    --verbose \\\n",
    "    --file $CUROUT/stuff_for_igv.tgz \\\n",
    "    --directory $CUROUT \\\n",
    "    $(basename $IGV_DIR)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Bash",
   "language": "bash",
   "name": "bash"
  },
  "language_info": {
   "codemirror_mode": "shell",
   "file_extension": ".sh",
   "mimetype": "text/x-sh",
   "name": "bash"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}