ImmunoSEQ parser fix

Software formats -S immunoseq and -S immunoseqv2 are now supported.

[experimental] Immunoseq format support

Use either -S immunoseq or -S immunoseqv3 when converting files to VDJtools format using Convert util. Which one to choose should be decided based on trial and error :)

Clonotype annotations patch 1

The patch adds additional (experimental) amino acid properties derived from PDB structures and fixes an issue with CalcCdrAAProfile

Clonotype annotations

What's new

• Allows storing clonotype-level annotations when working in VDJtools format, enabling better interoperability with other tools and extending analysis capabilities
• Annotations can be carried over from JoinSamples and PoolSamples result tables and VDJdb results (all these files are alredy in native VDJtools format)
• Added Annotate utility that allows to append various CDR3 structural (insert size, etc) and amino acid (e.g. GRAVY index) properties.
• Added parser for RTCR software

Improvements

• Original frequency in input sample is used in samples generated by VDJtools by default
• For certain software inputs renormalization (re-calculation of clonotype frequencies by dividing clonotype count by total read count of the sample) will be automatically performed upon importing data (Convert)
• Various *Filter, Decontaminate and Correct utils re-normalize frequencies. For *Filter routines --save-freq option can be used to output original sample
• PoolSamples can be applied to a single sample useful when one wants to work on amino acid CDR3 level, etc.

Fixes

• Fixed bug in SD calculation in CdrAaProfile module
• Implemented extensive testing by running various VDJtools modules on sample datasets in Travis CI

Known issues

• CdrAaProfile CLI fails with an exception, fixed in 1.1.1

Fixes for parser and R scripts

Fixes:

• Fixed bug in MiXCR parser (now VDJtools correctly handles duplicates in MiXCR full output) that can cause unexpected behavior in downstream analysis routines
• Fixed several bugs in R plotting to support latest ggplot version
• Fixed R script execution on Windows platform
• Fixed problem with spaces in vdjtools shell script

Additions and improvements:

• Added several new properties to CDR amino acid profile builder
• R scripts are now stored together with arguments so they can be modified and re-used
• Updated MacOS/Linux installation via Homebrew
• Add dedicated windows package with pre-installed R scripts

Vidjil parser, some improvements and fixes, Homebrew install

Improvements

• Allow to use Decontaminate based on read counts (old one was using clonotype fractions only)
• Proper normalization for Cdr3AAProfile routine in unweighted mode
• Add a parser for Vidjil software
• Add run script for compatibility with Homebrew (see our Tap)

Fixes

• Rare exception in Correct utility fixed
• ggplot2-2.x compatibility fix (guide keyword)

Fixes

• Fix V-J circus plot to correctly work in case of undefined segments
• Robust parser for MiGMAP
• Fix bugs & spelling in CDR3 amino acid feature module
• Add CDR3-center region for analyzing CDR3 amino acid features

API refactoring and fixes

• Substantial refactoring & javadocs for API classes (Java part)
• Code and resource cleanup
• Fixes for metadata operations
• Fix for new ggplot2 version

Metadata operations and compatibility fixes

New routines added

Add two new routines for metadata processing to simplify pipeline implementation:

• FilterMetadata filters metadata using an evaluated expression filter
• SplitMetadata splits metadata by column values

Compatibility

• VDJviz: metadata operations, sample-specific joining
• VDJdb: sample output
• Compatibility fix for MiGMAP 0.9.8 output

Minor

• Other fixes (rarefaction plotting, etc) and refactoring

Error correction and minor fixes/additions

• Added Correct module performing frequency-based error correction a-la MITCR
• Added an experimental option that computes sampling p-values for clonotype tracking
• Minor fixes