Msstitch

Changed
- Isobaric quant summarizing to features now throws out all PSMs with NA in a channel by default, keep them with --keep-psms-na-quant

Added
- Isobaric quant summarizing using denominators OR sweep OR median intensity
- Median normalization possible for isobaric quant
- Column in results "number of fully quanted PSMs", i.e. without missing values
- `msstitch deleteset` can delete a sample set from PSM tables and sqlite lookups
- `msstitch psmtable` can take `--oldpsms` to append new PSMs to an existing table (with identical search etc)

Fixed
- Bug (and free speed increase) where merged tables the following fields would be all wrong and all the same for each set: amount PSMs, amount peptides, amount unique peptides
- MS1 quant bug where sometimes wrong features where fetched to align to a scan

Added
- Added `--dinosaur` to store results from MS1 quant using dinosaur, include FWHM data
- Added median-intensity (i.e. not ratio) summarization to `isosummarize` and `proteins` etc

Changed
- Choose between MS1 apex and MS1 sum for quantification
- Choose between median and average when summarizing isobaric PSM data

Fixed
- Bug in MS1 feat to PSM scan aligment, `ORDER BY` had been removed in `9717325837853de159ede5e77191b0c7de032820`

Changed
- `msstitch merge --no-group-annotation` does output a protein(s) column for whatever accession is in the DB

Fixed
- Corrected typo in column to index on for ensg_proteins table (gene_id), which made merging ENSG very slow

Fixed
- Quant lookup on consensusXML stored channels in wrong order when >10 channels eg in TMT16
- --spectracol argument in msstitch psmtable did not work

Breaking
- Completely new interface, one command (msstitch), some subcommands, merged several
functionalities (e.g. SQLite creation/output) into fewer steps.

Added
- Possible to add ion mobility value of scans to PSM table

Changed
- Updated README
- Output column name changes (esp Gene ID/ Gene name from symbols)
- peptide sequence in first column of peptide table output
- trypsinize adds peptide index to fasta output as to not create duplicate entries
- PSM table creator does not longer take a mapfn from biomart, all info comes from fasta header
- Consequently, picked FDR can now work on ENSG ID and gene names
- SQLite creation for spectra has possiblity to specify output file
- Percolator output splitprotein different behaviour: can specify "known" prot headers which excludes all scans with those annotations,
when not specified it keeps any scan for a header (also if other headers are specified)

Fixed
- msslookup quant no longer stores duplicates if >500.000 scans, those first 500.000 were stored twice

Removed
- percolator split/merge/filter/qvality scoring, only kept protein identity splitting
- Protein error probability output