Bases2Fastq allows you to perform multiple different demultiplexing options after run completion, including demux from single or dual indexes, inclusion of UMI sequences into read headers, and other options like adjusting read filtering. Below are some common modifications to Bases2Fastq settings that can be added to your Bases2Fastq execution command. You can also visit the Bases2Fastq Online Help for additional details and more options.
QC Only
Using the –qc-only flag in your execution of Bases2Fastq will create a representative view of run metrics from a single tile of the sequencing run. The output will include the HTML report of QC metrics but does not generate any FASTQ files. QC-only mode allows for a much faster execution of Bases2Fastq that can be helpful for troubleshooting parameters, execution issues, or just to quickly obtain QC metrics for a run.
Run Bases2Fastq with an Alternate Run Manifest
You can run Bases2Fastq with a run manifest that differs from the one that was originally set for your sequencing run. This is useful if the original run manifest needs to be corrected, or if you need to demultiplex a pool that contains both single- and dual-indexed samples. To run Bases2Fastq with an alternate run manifest, add the –run-manifest flag and specify the location and filename of the alternate run manifest:
--run-manifest /input/alternate_manifest_filename.csv
Demultiplexing Pooled Single- and Dual-indexed Samples
If a mixed pool contains both single- and dual-indexed samples, Bases2Fastq can be run twice to ensure all samples in the pool have been captured during demultiplexing. Dual-indexed samples will be captured by using standard (default) Bases2Fastq settings. To capture single-indexed samples, the I1 mask will need to be set to N*. You can do this by either adjusting the I1 mask in the run manifest to N* (then use the above instructions to run Bases2Fastq with the alternate run manifest) or by using the –settings flag to overwrite settings in the existing manifest for the run:
--settings ‘I1Mask,I1:N*’
Single/Limited Tile Runs
Running Bases2Fastq on a single tile (or a few) from your sequencing run reduces computational time while still providing a reasonable representation of your sequencing data, which can be ideal for troubleshooting purposes. Unlike QC-only mode, specifying tiles in your Bases2Fastq execution will still create FASTQ files as part of the Bases2Fastq output. This is ideal if you want to change parameters and investigate the resulting FASTQ files. To run Bases2Fastq using a single tile, you will need to use both the --include-tile and --exclude-tile flags. For example, the following will demultiplex with a single tile from a run:
--exclude-tile L.R..C..S. --include-tile L1R10C01S1
Adjusting Filter Stringency and Placement
By default, Bases2Fastq uses the first 15 cycles of Read 1 and Read 2 for filtering, but this can be modified with the --filter-mask flag. The default filter-mask command is:
--filter-mask R1:Y15N*-R2:Y15N*
Generally, setting fewer cycles than the default will result in less stringent filtering, and more cycles in more stringent filtering. You can also change where the filtering occurs in the sequencing read. The filter-mask setting uses a base mask to define filtering cycles. You can find more details on this in the Base Mask online help. The example below shows an adjustment to the filter-mask that would apply the filter to cycles 2 – 16 of Read 1 and Read 2 (filtering is not applied to the first cycle of each read):
--filter-mask R1:N1Y15N*-R2:N1Y15N*
No Filtering
Running Bases2Fastq without any filtering can be useful for troubleshooting purposes, like calculating pass-filter (PF) rate for each sample. Demultiplexing without filtering includes reads that are often undesirable (e.g. originally filtered due to their low quality). However, comparing the PF reads for a standard (default) demultiplex with a no-filtering demultiplex can help identify libraries that are being preferentially filtered. See the Knowledge Base article on PF reads for more information. To run Bases2Fastq without filtering, add the following --filter-mask flag:
--filter-mask R1:N*-R2:N*
Single Read Filtering
You can also use the --filter-mask flag to filter by a single read only. This can be used for samples that were sequenced in a paired end format but only really require one read for analysis. It can also be useful if one read is of poor quality. If the quality of either Read 1 or Read 2 is very poor, that read and its pair will be filtered from the output for paired end sequencing by default, even if one of the associated paired reads was of high quality. If you would like to keep the high quality read from the pair, you can apply filtering to only one read, as in the example below:
--filter-mask R1:Y15N*-R2:N*
In the example above, default filtering settings are applied to Read 1, and no filtering is applied to Read 2. This will retain the high quality Read 1 reads associated with each polony that might have otherwise been filtered due to being paired with a poor quality Read 2. However, note that this does not improve or change the quality of the Read 2 reads, and they will now be included in the run output along with the Read 1 data. Adjusting the filter mask cycles outside of the default may impact quality scores for your sequencing run and should be used carefully. Additional QC, filtering, and removal of the included poor-quality reads will likely be needed prior to proceeding with other downstream analysis.