Editing Nanopore RNA Sequencing Protocol (section)

====Data Preparation====

Before we can begin, however, we need to extract some data from our raw <code>.fast5</code> files using ''Nanopolish'' and do some data formatting. First up, we need to index our raw reads using ''Nanopolish'', which can be done via the following:

<code>Bash</code>
<syntaxhighlight lang="Bash">
cd $NBASE
nanopolish index -d $NRAW/${SAMPLEID} ${SAMPLEID}_master.fastq
</syntaxhighlight>

Note that this should be called from the same directory as your .fastq master file. You can include the ''Guppy'' sequencing summary output by adding in the line <code>--sequencing-summary=sequencing_summary.txt</code> after <code>$NRAW/${SAMPLEID}</code> to speed up the indexing process, but I recommend against it if you get an error. In my experience, ''Guppy'' tells ''Nanopolish'' where files are located through this process, which may or may not be accurate.

Next, we need to generate the <code>eventalign</code> output from ''Nanopolish'', which aligns events from the raw data to the transcriptome for you. To do this, you need to indicate the location of your reference transcriptome file, your sorted and aligned <code>.bam</code> file, and your basecalled master <code>.fastq</code> file. The output of <code>eventalign</code> is '''''very large''''' with a good nanopore run (~23+ GB per chromosome of data), so I highly recommend considering mapping your reads to a transcriptome “subset” first and then running <code>eventalign</code> on that instead.

<code>Bash</code>
<syntaxhighlight lang="Bash">
nanopolish eventalign \
-r ${SAMPLEID}_master.fastq \
-b $NALI/${SAMPLEID}_aligned.bam \
-g $GREF/${GREFERENCEID} \
--scale-events \
--signal-index \
--threads 50 > ${SAMPLEID}_eventalign.txt
</syntaxhighlight>

Now we can do some ''m6ANet'' data prep and generate a few files:

<code>Bash</code>
<syntaxhighlight lang="Bash">
conda activate m6anet

m6anet-dataprep \
--eventalign $NBASE/${SAMPLEID}_eventalign.txt \
--out_dir $M6AO/${SAMPLEID} \
--n_processes 4
</syntaxhighlight>

The output of this will generate the following:

* <code>data.index</code>: Indexing of <code>data.json</code> to allow faster access to the file.
* <code>data.json</code>: json file containing the features to feed into ''m6ANet'' model for prediction.
* <code>data.log</code>: Log file containing all the transcripts that have been successfully pre-processed.
* <code>data.readcount</code>: File containing the number of reads for each <code>DRACH</code> positions in <code>eventalign.txt</code> document.
* <code>eventalign.index</code>: Index file created during <code>dataprep</code> to allow faster access of the ''Nanopolish'' <code>eventalign.txt</code> data during this step.