-
Notifications
You must be signed in to change notification settings - Fork 12
/
Copy pathNEWS
117 lines (86 loc) · 4.04 KB
/
NEWS
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
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
2018-04-16 v1.3.3
=================
Bug fixes
* handle the GSL dependency at run time, requiring it to be <= 1.16 or >= 2.3
2016-10-18 v1.3.2
=================
Bug fixes
* update tests to be compatible with Mac OS X
2015-08-10 v1.3.1
=================
Performance improvements
* implement the SQUAREM procedure to speed-um 'eqtlbma_hm' (thanks to C. Dai)
* reduce memory requirement of 'eqtlbma_bf' (thanks to G. Wang)
Bug fixes
* filter properly SNPs with low MAF when provided in "custom" format (thanks to
G. Wang)
* smoothly handle cases with not enough samples compare to the number of
predictors causing some permutations to fail (thanks to G. Wang)
2015-03-02 v1.3.0
=================
New features
* allow correlated errors when expression levels (some genes and/or some
individuals) are missing in some subgroups ("hybrid")
* in 'eqtlbma_bf', use --fiterr 0.5 by default compare to 0 before (motivated
by Wen, Biometrics), which can cause slitghly different values for the BFs
* keep median of log10(ABF) from permutations for QBF procedure (Wen, arXiv)
* add a tutorial with simulated data, see the manual
* allow to compute Bayes factors from summary stats (w/o covariance)
* major refactoring (use multiple libraries, handled by libtool) which causes
some options to change for eqtlbma_bf and eqtlbma_hm, see the manual
* add tests for the multivariate Bayes factor and the EM algorithm
* in 'eqtlbma_bf', skip SNPs with missing genotypes (with a warning)
* add 'eqtlbma_avg_bfs' to compute final quantities with more flexibility
* add 'utils_eqtlbma.R' in directory scripts/
* implement Poisson likelihood and quasi-likelihood for single-tissue analysis
but still beta (that is, users should not use it)
* add "type" model but still beta (that is, users should not use it)
Performance improvements
* use OpenMP to speed-up the EM algorithm fitting the hierarchical model
as well as to parallelize over SNPs in 'eqtlbma_bf' for permutations
* in 'eqtlbma_bf', handle tabix-indexed BED file for SNP coordinates
Bug fixes
* in 'eqtlbma_bf', add +1 to the start of gene coordinates from BED files
* in 'eqtlbma_hm', don't use underscore to split identifiers
2013-11-21 v.1.2.2
==================
Bug fixes
* minor buffer overflow in getDateTime() (Ryan Welch, from the University of
Michigan, raised the issue on GitHub)
2013-10-28 v.1.2.1
==================
New features
* minor changes in utils.cpp to allow compilation on Mac OS X (however the
bash scripts still require GNU getopt, not available by default on Mac)
2013-03-28 v1.2
===============
New features
* in 'launch_hm.bash', add option --reconf, --pi0 and --skip-ci
* in 'eqtlbma', change the format of the covariate files to make them readable by MatrixEQTL
* in 'eqtlbma', add option --outcv
* in 'eqtlbma', skip features (genes) if no coordinates are provided and remove option --ftr
* in 'launch_eqtlbma_parallel.bash', add option --snpD
* in 'eqtlbma', exit if an empty file of SNPs to keep is provided
* add all 'Makefile.in' to the repo so that users don't have to use 'autoreconf'
* add C++ code used to simulate data in Flutre et al.
* improve the manual, especially the FAQ
Bug fixes
* in 'hm', update pi0 in the hierarchical model (bug absent in paper's results)
* in 'eqtlbma', allow covariates only for a subset of the subgroups
* in 'eqtlbma', use the nb of covariates per subgroup for small sample size correction
* in 'eqtlbma', use 0.0 as log10(BF) of genes unexpressed in some subgroups
2013-02-08 v1.1
===============
New features
* add the 'hm' program implementing the EM algorithm to fit the hierarchical model
* in 'eqtlbma', put l10(BF) of absent genes for singleton configuration at 1 instead of NA
* in 'eqtlbma', add options --outss and --tricut
* in 'eqtlbma', report on stdout the running times using CLOCKS_PER_SEC
* for 'eqtlbma', add test for genes absent in some subgroups
Performance improvements
* in 'eqtlbma', avoid loading same genotype files for each subgroup
Bug fixes
* in 'eqtlbma', properly detect missing values when loading a VCF file
2013-01-16 v1.0
===============
first public release