FUCHS 0.3.1

1.1. Mates separate mapping (optional for PE data)

Note: the mate assignments should be consistent throughout the mapping and circular RNA detection process. In the following case, SamplePairedRead_1.fastq.gz is the first mate which also was the first mate in the STAR call.

# remap unmapped reads as single end to obtain double breakpoint fragments

$ gzip sample/Unmapped.out.mate1
$ mv sample/Unmapped.out.mate1.gz sample/Unmapped_out_mate1.fastq.gz
$ STAR --readFilesCommand zcat --runThreadN 18 --genomeDir [genome] --outSAMtype BAM SortedByCoordinate --readFilesIn [sample]/Unmapped_out_mate1.fastq.gz --outFileNamePrefix [sample].mate1.  --quantMode GeneCounts --genomeLoad NoSharedMemory --outReadsUnmapped Fastx --outSJfilterOverhangMin 15 15 15 15 --alignSJoverhangMin 15 --alignSJDBoverhangMin 10 --outFilterMultimapNmax 20 --outFilterScoreMin 1   --outFilterMismatchNmax 999 --outFilterMismatchNoverLmax 0.05 --outFilterMatchNminOverLread 0.7 --alignIntronMin 20 --alignIntronMax 1000000 --alignMatesGapMax 1000000  --chimSegmentMin 15  --chimScoreMin 15   --chimScoreSeparation 10  --chimJunctionOverhangMin 15 --twopassMode Basic --alignSoftClipAtReferenceEnds No --outSAMattributes NH HI AS nM NM MD jM jI XS  --sjdbGTFfile [annotation].gtf

$ gzip sample/Unmapped.out.mate2
$ mv sample/Unmapped.out.mate2.gz sample/Unmapped_out_mate2.fastq.gz
$ STAR --readFilesCommand zcat --runThreadN 18 --genomeDir [genome] --outSAMtype BAM SortedByCoordinate --readFilesIn [sample]/Unmapped_out_mate2.fastq.gz --outFileNamePrefix [sample].mate2.  --quantMode GeneCounts --genomeLoad NoSharedMemory --outReadsUnmapped Fastx --outSJfilterOverhangMin 15 15 15 15 --alignSJoverhangMin 15 --alignSJDBoverhangMin 10 --outFilterMultimapNmax 20 --outFilterScoreMin 1   --outFilterMismatchNmax 999 --outFilterMismatchNoverLmax 0.05 --outFilterMatchNminOverLread 0.7 --alignIntronMin 20 --alignIntronMax 1000000 --alignMatesGapMax 1000000  --chimSegmentMin 15  --chimScoreMin 15   --chimScoreSeparation 10  --chimJunctionOverhangMin 15 --twopassMode Basic --alignSoftClipAtReferenceEnds No --outSAMattributes NH HI AS nM NM MD jM jI XS  --sjdbGTFfile [annotation].gtf

Acquiring suitable GTF files for repeat masking

• It is strongly recommended to specify a repetitive region file in GTF format for filtering.

• A suitable file can for example be obtained through the UCSC table browser . After choosing the genome, a group like Repeats or Variation and Repeats has to be selected. For the track, we recommend to choose RepeatMasker together with Simple Repeats and combine the results afterwards.

• Note: the output file needs to comply with the GTF format specification. Additionally it may be the case that the names of chromosomes from different databases differ, e.g. 1 for chromosome 1 from ENSEMBL compared to chr1 for chromosome 1 from UCSC. Since the chromosome names are important for the correct functionality of DCC a sample command for converting the identifiers may be sed -i 's/^chr//g' your_repeat_file.gtf

Preparation of files containing the paths to required chimeric.out.junction files

• samplesheet file, containing the paths to the jointly mapped chimeric.out.junction files

$ cat samplesheet
/path/to/STAR/sample/joint_mapping/chimeric.out.junction

• mate1 file, containing the paths to chimeric.out.junction files of the separately mapped first read of paired-end data

$ cat mate2
/path/to/STAR/sample.mate1/joint_mapping/chimeric.out.junction

• mate2 file, containing the paths to chimeric.out.junction files of the separately mapped first read of paired-end data

$ cat mate2
/path/to/STAR/sample.mate2/joint_mapping/chimeric.out.junction

Running DCC

After performing all preparation steps DCC can now be started:

# Call DCC to detect circRNAs, using HEK293 data as example.

$ DCC @samplesheet \ # @ is generally used to specify a file name
      -mt1 @mate1 \ # mate1 file containing the mate1 independently mapped chimeric.junction.out files
      -mt2 @mate2 \ # mate2 file containing the mate1 independently mapped chimeric.junction.out files
      -D \ # run in circular RNA detection mode
      -R [Repeats].gtf \ # regions in this GTF file are masked from circular RNA detection
      -an [Annotation].gtf \ # annotation is used to assign gene names to known transcripts
      -Pi \ # run in paired independent mode, i.e. use -mt1 and -mt2
      -F \ # filter the circular RNA candidate regions
      -M \ # filter out candidates from mitochondrial chromosomes
      -Nr 2 2 \ minimum number of replicates the candidate is showing in [1] and minimum count in the replicate [2]
      -fg \ # candidates are not allowed to span more than one gene
      -G \ # also run host gene expression
      -A [Reference].fa \ # name of the fasta genome reference file; must be indexed, i.e. a .fai file must be present

# For details on the parameters please refer to the help page of DCC:
$ DCC -h

Notes:

• By default, DCC assumes that the data is stranded. For non-stranded data the -N flag should be used.

• Although not mandatory, we strongly recommend to the -F filtering step

Output files generated by DCC

The output of DCC consists of the following four files: CircRNACount, CircCoordinates, LinearCount and CircSkipJunctions.

• CircRNACount: a table containing read counts for circRNAs detected. First three columns are chr, circRNA start, circRNA end. From fourth column on are the circRNA read counts, one sample per column, shown in the order given in your samplesheet.

• CircCoordinates: circular RNA annotations in BED format. The columns are chr, start, end, genename, junctiontype (based on STAR; 0: non-canonical; 1: GT/AG, 2: CT/AC, 3: GC/AG, 4: CT/GC, 5: AT/AC, 6: GT/AT), strand, circRNA region (startregion-endregion), overall regions (the genomic features circRNA coordinates interval covers).

• LinearCount: host gene expression count table, same setup with CircRNACount file.

• CircSkipJunctions: circSkip junctions. The first three columns are the same as in LinearCount/CircRNACount, the following columns represent the circSkip junctions found for each sample. circSkip junctions are given as chr:start-end:count, e.g. chr1:1787-6949:10. It is possible that for one circRNA multiple circSkip junctions are found due to the fact the the circular RNA may arise from different isoforms. In this case, multiple circSkip junctions are delimited with semicolon. A 0 implies that no circSkip junctions have been found for this circRNA.