-
Notifications
You must be signed in to change notification settings - Fork 0
/
CHANGELOG
78 lines (58 loc) · 4.09 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
76
77
78
== Version 2.1.0 ==
New Commands:
1. Added a "load-categorical" command for loading categorical covariates.
New Features:
1. Added a "Number of Cases" column in the case of Logistic Regression.
2. Added higher resolution near 0 for displayed p-values in regression methods.
3. Added a "Model Permuted P-value" column for permuted regression methods
4. Added the ability to report (and correct for) the genomic inflation factor, by 2 different methods.
5. Added a "Rank Permuted P-value" column for permuted regression methods.
6. Added sample pre-filtering to loading of trait and categorical files
7. Added an option to use Firth correction in logistic regression
8. Added the ability to get the gender from a FAM file
Bug Fixes:
1. load-trait can now account for files with Windows line endings (CR+LF vs. LF)
2. Fixed almost certain deadlocking when using the "--one-sided" option in a regression method
3. Fixed deadlocking when using Weighted encoding + MPI
4. Fixed potential deadlock when using MPI + threading
5. Fixed incorrect reporting of p-values when using single-variable weighted encoding + covariates
6. Allow the recode-alleles command to give an alternate allele when only one allele is present
== Version 2.0.0 ==
Completely rewrote PLATO, changing many options and changing the way it runs.
== Version 1.2.2 ==
Bug Fixes:
1. Fixed error in reading plink binary format files for case where the files were not in chromosomal position order.
== Version 1.2.1 ==
New Feature:
1. Added GXE regression testing in interaction method.
Bug Fixes:
1. Fixed memory leak in logistic regression.
2. Fixed crash when using named phenotypes from traits file in logistic regression.
3. Output columns in logistic regression now align propery with headers.
Compilation Fix:
1. Changed macro for malloc.h inclusion to HAVE_MALLOC_H for compiling on systems without malloc.h ( such as Macs ).
== Version 1.2 ==
New Features:
1. Added -update-ids command line option. Takes an input file of the format: OldFamID OldIndID NewFamID NewIndID The ids are updated prior to performing a formatting step (output-ped, output-bin, etc
2. User can specify covariates to include in linear regression with the -covars-name and -covars-number parameters.
3. MDROutput process can now use alternative phenotype provided by trait-file
4 Added a regression interaction analysis. It can be accessed using the batch command 'interaction' and can accept a bio-snp-file as epistasis does to run pairwise regression analyses.
Bug Fixes:
1. In FST calculation checks that the number of groups specified doesn't equal the number of samples.
2. Replaced code to filter output based on -epi-alpha1 in Epistasis module.
3. Added markerGreater(...) method to be used as comparator in InputFilter class sorting. This was a fix for an issue with -excMarkers flag not excluding all markers in the locus input file.
4. Fixed ExcludeLocusFilter(...) method in InputFilter class to make sure all markers in the input locus file are excluded regardless of order.
5. Changed to catch MethodException reference so messages are displayed correctly.
6. In Sample::addMicroSat and MarkerGenoEff::calcOne, changed map access from .at to [] because .at throws an exception when key doesn't exist and in this case, we are adding a new key/value pair to the map
7. Fixed Helpers::removeBeginWhiteSpace to properly remove whitespace at beginning of string
8. Added warning to BINOutput when trying to include a microsatellite marker in a plink output file
9. For LGEN file input, fixed missing data problem so that any marker not appearing in the input file would be set as missing for processing.
10. For Eigenstrat output, corrected 0 and 2 output for alleles in the ancestry map file option.
Output:
1. Linear Regression now outputs the number of covariates included in the model
Style/Format/Other:
1. Many lines of commented out code removed.
2. Moved opening braces to separate lines
3. Changed STL map and vector access from [] operator to .at function.
4. Cast size() function of vectors as ints to remove compiler warnings
5. Removed R dependency and MARS filtering techniques