Chimeric transcript detection

Chimeric reads covering two predicted transcripts were extracted using the ChimerID scripts (Parker et al., 2020) (https://github.com/bartongroup/Simpson_Barton_Nanopore_1, Farhat, 2021a copy archived at swh:1:rev:1b509454a9e25a8c81be5092f8e525ca00e7b5a5) from native RNA Nanopore aligned reads within a Anaconda environment containing the following packages (snakemake ;pysam ; samtools ; scipy ; statsmodels ; py-bgzip ; tabix ; pytables ; bedtools ; minimap2 ; panda=0.25.3). Prior to running the analysis, the ToxoDB gene file (in gff format) was processed by AGAT (https://github.com/NBISweden/AGAT, Farhat, 2021b copy archived at swh:1:rev:692791aa30d253bd2ff83397d0dc1a74b5b52adb) to convert to GTF and modify the format. For practical reasons, all KEXXXX contigs were removed from the GTF and genome fasta file. Once extracted, chimeric reads were counted against transcripts in the TGME49 gtf file using Htseq-count. This count file was also supplemented with normal counts of standard transcripts. Data was then normalized using a total aligned read normalization and ratios of Chimera over Total reads were calculated for every transcript in every condition. Log2Fold change in these ratios were then calculated to compare UT against IAA conditions for every KD condition (CPSF4, METTL3, and MORC). Transcripts with readthrough counts below three were excluded from further analysis. Of note, we had no mAID UT condition for MORC so we used the METTL3 UT condition as a comparison to calculate Log2Fch.