forked from laurentnoe/yass
-
Notifications
You must be signed in to change notification settings - Fork 0
/
NEWS
163 lines (148 loc) · 8.19 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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
Yass-1.01 [dec 2002]
- first public version in december 2002 : in fact, i started this program,
together with many other things that were never finished, during my
master internship that started around march/april 2002... good
old time :-)
Yass-1.02 [september 2003]
- output entropy filter on alignments
- transition constrainted spaced seeds added
- alignments methods revised (alignement speed ~ x1.3)
- hit criteria revised (speed ~ x1.2)
- many bugs solved
Yass-1.03 [october 2003]
- BLAST tabular like output
- multifasta support with file selection (-S)
- more scoring matrices
- less global variables ;-)
- autoconf format added (that's helpfull)
Yass-1.04 [november 2003]
- -r 2 option changed
- bug with missed alignments solved (alignandfree take the MAX now)
- -T option only affect direct repeats on one single file ... (no
palindromic ones)
- align algorithm changed (adequate greedy selection made more efficiently)
Yass-1.05 [august 2004]
- Blast tabular output / header
- Matrix cost options
- Lambda/K value can be changed by hand.
- Xdrop modifications / Expectation value to cut off
- Seed has changed (optimised on random sequences / with constrain on
jokers and transitions free elements)
- several bugs solved (Thanks, Martin !!)
- problem with W,S letters partially solved (replacement)
- first version of the web interface was done by Ougas ELMI HOUSSEIN [during summer internship]
Yass-1.06 [november 2004]
- -W option for grouping DNA fragmented alignements
[ work done by Steven CORROY (in only two months during summer 2004!! good job),
- some specifics bugs have been solved since.
- several Warnings and optimizations for the -W option.
- improved methods both in implementation and also definition.
]
- optimized alignment between seeds and greedy one over seeds (now i have to optimize the core "assemble" part).
- -O0 non optimization bug removed (strange one, isnt it ?)
- "||||15120" number on alignment display removed
- random walk polynom computed now with qpow (faster, of course ;-) )
Yass-1.07 (work done by Christophe VALMIR) [february 2005]
- Cache efficient list structure
- Add Diagonal hash table
- One multi fasta BUG solved
Yass-1.08 (work done by Christophe VALMIR) [july 2005]
- Extension MA BUG solved
- Regroup BUG solved
- Lambda and K computed cordance with sequences selected
- Make -C option for modifed scoring matrices
- Yass used a multi-seed
- The substitution matrix was changed (default:-M 3)
- Bug for compiling MAC solved
Yass-1.09 (work done by Christophe VALMIR) [october 2005]
- thread for the -r 2 option (only with "configure --with-thread").
- -r 2 option is the default option
- Bug for scoring solved.
- More regrouped.
Yass-1.10 (back to LN) [october 2005 - december 2006]
- cosmetic and optimisation work (debugging of -r 2 )
- redefinition of several functions
- key-compress bug solved
(now all these options can be selected from the util.h)
- one-multifasta bug solved
- several macro redefined
- extention algorithms updated
- memory check option for standalone version
- statitics for each thread
- More regrouping.
Yass-1.11 [january 2007 - april 2007]
- lowercase regions parameter
- unusual letters complements bug solved
- entropy on spaced words ...
- output sorted according to 4 criteria (+ 4 )
- new parsing script
- due to a general demand, i have changed the "." and ":", now they
represent tranversions "." and transition ":" repectively
- key only composed of pairs of amino acids are removed
- MEMOPT type bug solved (i forget some copy)
Yass-1.12 [sept 2007 - december 2007]
- single hit added as an option ... group size has been removed.
- low complexity now does not interfer between threads ... global
variables were shared before.
- Gumbel Law Lambda and K are computed with full consideration of the
background nucleotide distribution for both sequences ...
- bug solved when left alignment was computed : some unreleavants
positions were added.
- bug solved when runs of 'N' were complemented.
- bug solved on multiseed span linking
- bug solved on the rightmost alignments
- bug solved on initial -M scoring parameter
Yass-1.13
"Houston, we've had a problem here"
Yass-1.14 (* : work done by Antoine DE MONTE) [january 2008 - january 2013]
* multifasta "full support" (for the first file too, with the "-S 0" parameter)
* multithreading improved, parametrized for the multifasta first file
* 64 bits support (files larger than 2Gb)
- PSL output format added
- bug solved on multiseed span linking
- bug solved on the rightmost alignments
- several bugs solved on the alignment_SG_Border part
- bug solved on the min_max_span of the default seed (inverted ?? does not appear in previous versions)
- bug on the "mask copy" (RedMask function)
- new implementation of the greedy "align tuples" functions
- bug on "-c 1" tuple size (was corrected during extention ... now solved during hit)
- many improvements in the "align.c" main method x 2 ... work in progress on this part ...
- most of the alignments are now provided by the heuristic "Border" functions (banded alignment) + 2 bugs solved on it
- output bug (on one case, indels were numbered ...)
- sort on %identity (query/text/alignment) added
- hash cache table removed (not used)
- parser modified (bug on inner '>' when fasta file is incorrectly formatted) + memory on header kept
- bug on single file extention (when following the main diagonal)
- autotools "-lm" option changed from LDFLAGS to LIBS (does not change the compiled version ...)
- heuritic "improved" on resolve_overlaps "hits" (more DP-Alignment is done when this case occurs :
this do not apply however on non resolve_overlaps "hits" that are thus "seed pattern dependent".
Yass-1.15 : [january 2013 - april 2020]
- Small improvement on first file (query) index on multifasta : avoid the "more than quadratic time" (due to reuse of kword first index)
it is better than before, but no still perfect has it does not build index on multifasta file into a single index.
- dereferencing pointer compiler warning fixed
- bug on the index "imask" overflow fixed (not detected before : compiler issue ?)
- small bug on bad Entropy value (small difference) fixed (found by Aude Darrack), fixed in 1.15 [november 2014]
- not a bug: some warning added on the seed pattern when bad ones are in the input
- First hit (tuple) selection [Ongoing]
- Disabling the assemble_align thread at very first time, or for small sequences [august 2015]
- Fixing a regroup scoring evaluation (optimistic way with overlaps) + a regroup window increment
- Patching a Resolve_Overlap annoying Bug [april 2016]
- Patching a Lazy Alignment large indel Bug + some valgrind [may 2016]
- Changing the behavior of -s 6x to avoid redudancy with -s 4x [august 2016]
Now similar to -s 7x (BLAST), with the difference that the queries blocks
are not kept in the same order as input, but ordered by their best overall score ...
- Solving some memory leak, uninitialized value, and a bug in "regroup.c": cleaning on customer but not on producer [december 2016]
- Probability stats computed on forward/reverse strand [february 2017]
- Hotfix on regroup [july 2017]
- Line number report [february 2018]
- Non-indexing errors have been replaced by warnings [july 2018]
– Correct penalty for negative overlap between two alignments [september 2018]
- Separating the "MAminscore" (evolving when too many alignemnts) and the "TUminscore" (to start an alignment) [september 2018]
- When saturation occurs, updating the "MAminscore" to remove alignments that were previously computed, but are under the saturation filter, and filter it before regroup occurs to get homogeneous results [january 2019]
- When grouping selction occurs, use the main queue in both directions, with a "last element" trick to search inside the main queue (does not change the algorithm, but its complexity) [march 2019]
- Score used to upgrade tuples was reconsidered and set to its worst case [march 2019]
- yass2dotplot.php added for Offline Dotplot Work [april 2020]
Yass-1.16 : [march 2022 - ...]
- MoW (Minimally overlapping words) seeds strategy "sampling" added to speed up search on large sequences [march 2022]
- yass2dotplot.php warnings, yass2blast.pl shebbang [mars 2024]
- yass2dotplot.php svg additional output and parameter [may 2024]