-
Notifications
You must be signed in to change notification settings - Fork 32
/
Copy pathChangelog
75 lines (70 loc) · 3.71 KB
/
Changelog
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
2020-06-15 Changes in 2.1.5 Andreas Wilm <[email protected]>
* Fixed a long-standing Heisenbug related to indel formatting (closes #89)
2020-01-04 Changes in 2.1.4 Andreas Wilm <[email protected]>
* Now using htslib API thanks to John Marshall. Requires htslib >= 1.4. New configure flags
replacing SAMTOOLS and HTSLIB: --with-htslib
* Minor fixes in parallel caller closing #85
* Fixed #58
* Enhancements to lofreq2_add_fake_gt.py
2017-06-21 Changes in 2.1.3.1 Andreas Wilm <[email protected]>
* Fixed bug introduced last minute in 2.1.3 that creates segfault if call is used without -o
2017-06-21 Changes in 2.1.3 Andreas Wilm <[email protected]>
* Maintenance release before major rewrite
* Added Python3 support
* Added best practices snakemake workflow
* Little easier on memory in high coverage situations
* Added --force-overwrite option to 'call'
2015-05-19 Changes in 2.1.2 Andreas Wilm <[email protected]>
* 'indelqual' now allows to read bam from stdin
* Fixed bug in 'call' which resulted in negative phred quality
filter, when pvalue alpha was above 1 and number of tests was low
* 'indelqual' dindel now deletes BI/BD before inserting
* remove unnecessary dependency on kaln.h (not present in samtools
1.2)
* 'uniq' now closing output vcf filehandle on error, thus always
writing at least a header (reported by DNANexus)
* Added HRUN info field to output vcf
* Fixed calling of indel consvars
* Removed options (and use of) cons-as-ref and skip-n. now reference
is always used by default to call against and n's are always
skipped. also means the consensus variants (CONSVAR) concept
disappeared
* Set DEFAULT_MIN_PLP_IDQ to zero
* Caught yet another variant of the reference sequence name
mismatch problem
* 'viterbi': memory allocation now mainly dynamic.
fixes observed segfault on pacbio reads (unclear why though)
* Low AF false positive multi-allelic 1bp indel adjacent to
poly-AT now filtered by default.
* 'indelqual': added support for adding uniform insertion and
deletion qualities (instead of just indel qualities)
* indel calling: fixed index violation while accessing pdi[u] in
idaq happening while processing pacbio reads. added bound check as
hack (idaq() mostly illumina specific anyway)
* Removed MAX_READ_LEN globally
* 'call': added special case for SB test: if ref is entirely missing
and we have alts on only one strand fisher's exact test will
return 0, which is most certainly not what we want. setting to
INT_MAX instead
* vcfset: only-[type] now correctly dealt with in vcf2 on top of vcf1.
* vcfset: fixed bug which match vars even if they only overlapped
partially (now also checking position instead of relying on
tabix iterator)
* Reference sequences now converted to uppercase after fetching to be safe.
This also addresses the "AQ-bug" where low AQ values were reported
because of a lower-case reference
* 'pparallel': made bed reading function standard conform and more
fault tolerant. e.g. now allowing browser and track lines
* 'somatic': now also producing germline indels
* 'somatic': min cov lowered to 7
* 'somatic': normal stringent now has separate parameters, i.e. independent of
tumor stringent (set to fdr 1%)
* 'somatic': sq ignore normal now ignoring indels and snvs
* 'somatic': Added support for multiple ignore vcf files
* 'filter': corrected wrong info about default sb mtc method
* Changed MQ0 prob to a generic 0.5
* Fixed bug in fdr application in filter: previous versions
called more tests significant than actually true. Fixed by setting
all values to not significant right after calling fdr()
* 'somatic': added --min-cov option
Changes in versions before 2.1.2: See http://csb5.github.io/lofreq/blog/