# Changelog

## 3.6.3
* Documentation for processing demultiplexed FASTQ files

## 3.6.2
* Additional validation of FASTQ records to detect badly concatenated files
* Documentation for read ID check policy

## 3.6.1
* Add a specialized template barcode policy to optimize a common use case, where a template
  matches 2 barcode regions, each of which has a short prefix, separated by a fixed length
  run of arbitrary barcodes.

## 3.6.0
* Support reading multiple reads files sequentially as part of a single run

## 3.5.0
Released as open source under a BSD 3-Clause license

## 3.4.4
No user-facing changes

## 3.4.3
## Improvements
* Report empty conditions as "Unlabeled Sample Barcodes"

## 3.4.2
## Improvements
* Add support for different read ID checking policies, including one specific to Illumina which
  ignores divergent material after the first space in the ID line.
## Bugfixes
* Corrected fixed barcode prefix policies to allow the 'FIXED@' syntax, per the documentation;
  previous versions required the syntax 'FIXED:'. For backwards compatibility, the former syntax is
  still supported but in future versions this may be removed.

## 3.4.1
No user-facing changes

## 3.4.0
## Improvements
* Added support for reading row barcodes from paired-end sequencing data

## 3.3.4
## Improvements
* Improved error message in the case where reads are too short

## 3.3.3
No user-facing changes

## 3.3.2
No user-facing changes

## 3.3.1
## Improvements
* Advanced setting `--always-count-col-barcodes` (disabled by default). When enabled, PoolQ counts
column barcodes from reads even if a row barcode is not found, which changes the per column barcode
metrics `Matched Sample Barcode` and `% Match`, reported in the quality file.

## 3.3.0
## Improvements
* Support UMI barcodes

## 3.2.11
## Improvements
* Allow empty barcode IDs in row and column reference files

## 3.2.10
No user-facing changes

## 3.2.9
## Improvements
* Updated manual

## 3.2.8
No user-facing changes

## 3.2.7
No user-facing changes

## 3.2.6
No user-facing changes

## 3.2.5
No user-facing changes

## 3.2.4
No user-facing changes

## 3.2.3
### Bugfixes
* Correctly skip writing correlation file when there is only a single construct barcode or a single
  experimental condition

## 3.2.1
### Improvements
* Added whitespace to run info file for readability

## 3.2.0
### Improvements
* Add `Normalized Match` column to quality file
* Change quality file header from `Read counts for sample barcodes without associated conditions:` to
  `Read counts for most common sample barcodes without associated conditions:`
* Change terminology from `scores` to `counts`
* Log version and run configuration in run info file
* Add changelog

## 3.1.1
### Bugfixes
* Correctly skip writing correlation file when there is only a single experimental condition

## 3.1.0
No user-facing changes

## 3.0.5
### Bugfixes
* Fixed path to PoolQ jar in `Makefile` and scripts distributed with PoolQ
### Improvements
* Added regression tests for barcode counts and log-normalized counts files

## 3.0.4
### Bugfixes
* Fixed delimiter in `Construct IDs` field of log-normalized counts files

## 3.0.3
No user-facing changes

## 3.0.2
No user-facing changes

## 3.0.1
### Improvements
* Made 5' search limit optional when specifying a row barcode search template

## 3.0.0
* Complete re-write of PoolQ