New feature: Downgrade errors

It is now possible to downgrade errors to warnings for specific handlers. Prior to evaluating the QC criteria, the config will be updated so that warning threshold = error threshold and error threshold = unknown. Read more about usage in the docs.

New feature: Index white-list

It is now possible to specify white-list indexes that only cause warnings even if they occur above the significance level.

Small QC config update

Patch:
Updated Q30 threshold for hiseqx_v2.

QC critera for NovaSeq SP 101 bp

Patch:
QC criteria for NovaSeq SP, read length 101 bp, have been added to the default config.

More information on dual indexes and errors instead of warnings for significant indexes

This release changes the behavior when significant unidentified indexes are found from warning to giving an error, i.e. it will make the exit flag of the program 1 instead of 0.

In addition to this it introduces more sophisticated checks for dual indexes. We will now check each index in a dual index combination separately to see if one of the two e.g. has been reversed, or if the found index matches the complement of an index in the samplesheet.

Bug fix: Fixed missing version bump

v3.2.1

Bump version to 3.2.1

Support for NovaSeq SP flowcells

This release include the following updates:

• The readme now refer to Python 3.6 instead of Python 3.5 since the later version is needed for some modules to run properly.
• Support and QC criteria for NovaSeq SP is now added.
• QC criteria and read length for some instrument types are now in the configuration file.

Bug fix: iSeq Instrument ID correction

The documentation regarding iSeq naming conventions was incorrect, the naming convention for iSeqs should be that they start with FS. This is fixed in this version.

Index checking

This release introduces the UnidentifiedIndexHandler. This will aid in identifying indexes which end up as unidentified after demultiplexing. This is a nifty feature since often when there is a high amount of unidentified reads, this is cased by errors in the samplesheet. For example, when using dual indexes index one and index two may have swapped places.

The new handler will look for in indexes found in undetermined at a frequency compared to the number of reads on the lane above what is specified by significance_threshold (default: 1%) in the config and compare them to the samplesheet to decide if:

• dual indexes have been swapped
• the index has been reversed
• the index is the reverse complement
• the index is the complementary index
• the index is present in another lane

An example of that the output you will get could look like is:

WARNING  QC warning: Index: TTCGAT on lane: 8 was significantly over represented (10.0%) at significance threshold of: 1%.
WARNING  QC warning: 	We found a possible match for the reverse of tag: TTCGAT, on: Lane: 2, for sample: SampleA. The tag we found was: TAGCTT

The first line will warn that the index is represented over the configured threshold, and the second line will suggest what might be wrong (based on the scenarios listed above).

Bug fix: Fix configuration for iSeq

v3.0.1

Bump version 3.0.1