yass Code
Brought to you by:
ayemoy
Menu
▾
▴
Tree [7d8633] master / History
noe
noe
Laurent Noé
Read Me
* YASS
* Similarity search in DNA sequences
* version 1.15
** LICENSE
YASS is distributed under the dual-licence of the CeCILL (version 2 or any later
version), and the GPL (version 2 or any later version), so you can use, modify
and redistribute the program under these licences with almost no restriction.
** COMMAND SYNTAX
* Usage :
yass [options] { file.mfas | file1.mfas file2.mfas }
-h display this Help screen
-d <int> 0 : Display alignment positions (kept for compatibility)
1 : Display alignment positions + alignments + stats (default)
2 : Display blast-like tabular output
3 : Display light tabular output (better for post-processing)
4 : Display BED file output
5 : Display PSL file output
-r <int> 0 : process forward (query) strand
1 : process Reverse complement strand
2 : process both forward and Reverse complement strands (default)
-o <str> Output file
-l mask Lowercase regions (seed algorithm only)
-s <int> Sort according to
0 : alignment scores
1 : entropy
2 : mutual information (experimental)
3 : both entropy and score
4 : positions on the 1st file
5 : positions on the 2nd file
6 : alignment % id
7 : 1st file sequence % id
8 : 2nd file sequence % id
10-18 : (0-8) + sort by "first fasta-chunk order"
20-28 : (0-8) + sort by "second fasta-chunk order"
30-38 : (0-8) + sort by "both first/second chunks order"
40-48 : equivalent to (10-18) where "best score for fst fasta-chunk" replaces "..."
50-58 : equivalent to (20-28) where "best score for snd fasta-chunk" replaces "..."
60-68 : equivalent to (70-78), where "sort by best score of fst fasta-chunk" replaces "..."
70-78 : BLAST-like behavior : "keep fst fasta-chunk order", then trigger all hits for all good snd fasta-chunks
80-88 : equivalent to (30-38) where "best score for fst/snd fasta-chunk"" replaces "..."
-v display the current Version
-M <int> select a scoring Matrix (default 3):
[Match,Transversion,Transition],(Gopen,Gext)
0 : [ 1, -3, -2],( -8, -2) 1 : [ 2, -3, -2],(-12, -4)
2 : [ 3, -3, -2],(-16, -4) 3 : [ 5, -4, -3],(-16, -4)
4 : [ 5, -4, -2],(-16, -4)
-C <int>[,<int>[,<int>[,<int>]]]
reset match/mismatch/transistion/other Costs (penalties)
you can also give the 16 values of matrix (ACGT order)
-G <int>,<int> reset Gap opening/extension penalties
-L <real>,<real> reset Lambda and K parameters of Gumbel law
-X <int> Xdrop threshold score (default 25)
-E <int> E-value threshold (default 10)
-e <real> low complexity filter :
minimal allowed Entropy of trinucleotide distribution
ranging between 0 (no filter) and 6 (default 2.80)
-O <int> memory limit of the number of ungapped alignments (default 1000000)
-S <int> Select sequence from the first multi-fasta file (default 0)
* use 0 to select the full first multi-fasta file
-T <int> forbid aligning too close regions (e.g. Tandem repeats)
valid for single sequence comparison only (default 16 bp)
-p <str> seed Pattern(s)
* use '#' for match
* use '@' for match or transition
* use '-' or '_' for joker
* use ',' for seed separator (max: 32 seeds)
- example with one seed :
yass file.fas -p "#@#--##--#-##@#"
- example with two complementary seeds :
yass file.fas -p "##-#-#@#-##@-##,##@#--@--##-#--###"
(default "###-#@-##@##,###--#-#--#-###")
-c <int> seed hit Criterion : 1 or 2 seeds to consider a hit (default 2)
-t <real> Trim out over-represented seeds codes
ranging between 0.0 (no trim) and +inf (default 0.001)
-a <int> statistical tolerance Alpha (%) (default 5%)
-i <int> Indel rate (%) (default 8%)
-m <int> Mutation rate (%) (default 25%)
-W <int>,<int> Window <min,max> range for post-processing and grouping
alignments (default <64,65536>)
-w <real> Window size coefficient for post-processing and grouping
alignments (default 16)
NOTE : -w 0 disables post-processing
Scoring System :
-M <int> : choose a preselected matrix and gap opening and extension penalty
-----------------------------------------------------------------------------
the default "-M 3" gives fowoling scores :
+5 for match reward,
-4 for transversion penalty,
-2 for transition penalty,
and
-16 for gap opening penalty.
-4 for gap extention penalty.
but you can also choose the
-C <int>(,<int>(,<int>(,<int>))) : choose a scoring system
-----------------------------------------------------------------------------
if you give 2 parameters : match reward, mismatch penalty
3 parameters : match reward, transversion penalty,
transition penalty
4 parameters : match reward, transversion penalty,
transition penalty,
other penalty (non ACGTU letters)
-G <int>,<int> : choose a gap opening and extension penalty
------------------------------------------------------------------------------
for gap opening penalty, gap extention penalty.
-d <int> : choose Display preferences
-------------------------------------
-d 0 gives maximal positions of alignments
-d 1 gives ... + statistical parameters and quick alignments
(it means that those alignments are computed using seeds
as anchors, which is not always the best solution, but
the fastest one)
-d 2 blast like tabular output
-d 3 "position - size" tiny display
-s <int> : choose Sorting
-------------------------
it is possible to sort according to score (-s 1), positions on
the first (-s 4) or the second file (-s 5), and also, for
multifasta files to sort according to the first file chunks (-s 10
"to" 15), the secongd file chunks (-s 20 "to" 25) or both of them
(-s 30 "to" 35)
-S <int> : choose the first multifasta chunk
--------------------------------------------
this parameter is by default 1, since it selects the first chunk
of the first multifasta file : you can choose any valid chunk or
either fix it to "0" if you want to treat al ofl them
-p <pattern> : modify the seed Pattern
--------------------------------------
This element represents the seed pattern to be searched
* examples of seed patterns:
- PatternHunter : ###__##_#__###
- Mandala : ###__#_##_### (forcing small span)
####_##_###
- YASS : #@##_#__##__#@# (on Bernoulli
model with constrains on seed jokers/tr-free elements)
##@_#@#__#_### (on pure Bernoulli model)
- 2 seeds :
some examples to detect :
- very small random regions :
###-###-####,###-#--#--#-####
- small random regions :
##-#----###-####,###--#-#-#--#--###
- random regions :
##-#-##---##-##--#,##-###--#---#----###
- large random regions :
##--#---#-#--###---##,##-#--#-#--#--##-##
- very small transition rich random regions :
###@##-##-#@#,###-@-#--@--#####
- small transition rich random regions :
##@#-#-##@-###,#-##--#--#@#--@-###
- transition rich random regions :
###-#----#-#--#@#-@#,##-#--@#-#-###--@#
- large transition rich random regions :
#@#--#----###--@-##,#-#-#@---#-#-@-##-##
other examples :
mixed regions (chr-inver-fungi) :
#@-##-##-#--@-###,#-##@---##---#---#@#-#
mixed regions (chr-best) :
#-###@#--#-#-@##,###@-#-#------@#-#--##
mixed regions (chr-avg) :
##-#-#@#-#--@###,#-##@--#----#--##@##
-e : alignment Entropy
-----------------------
YASS does a filtering step based on the triplet composition
of the alignment
-e 0.0 : filter is off.
-e 3.10 : filter is set to its default level
-e 5.99 : filter is set to its highest level (don't use this)
PS : this option is subject to changes in future versions
-r <int> : reverse strain considered
------------------------------------
YASS considers both forward and complementary reverse strain
of the first sequence if -r 2 set. You can select with -r 0 only
direct repeats or only complementary inverted ones with -r 1
(default is both forward and complementary : -r 2 )
-S <int> Select query sequence from a multifasta file (default 1)
----------------------------------------------------------------
If the first file provided is not fasta but multifasta, you can choose
(using this parameter) which fasta sequence will be the query (by
default the first one).
-T <int> "anti Tandem trick"
------------------------------
Forbid aligning too close regions (e.g. Tandem repeats)
valid only for comparing a single-sequence file against itself
-a -i -m : seed grouping parameters
-----------------------------------
YASS groups seeds before extension and then extends them once grouped.
These parameters enable you to modify the grouping criteria and extention
limits that are fixed (before post processing).
"-m" gives the expected mutation rate (25%) : it can be increased to 30%
or 40% but no more in practice.
"-i" gives the expected indel rate (8%) : can be also modified to ~ 10%
"-a" give the bound so 95% of (when fixed to 5%) of the distribution of
indels/mutations between seeds are captured during the extention process.
-W <min,max> -w <mul> : Post-processing parameters
--------------------------------------------------
In order to group some consecutive alignments into a better
scoring one, post processing tries to group neighbor alignments in
an iterative process : by applying several time a sliding windows
on the text and estimating score of possible groups formed.
Windows size can be controlled according to a geometric pattern
<mul> and two bounds <min,max>.
-w 0 (or 1) disables this post-processing step.
References
==========
YASS is presented in the following paper. If you use YASS, please cite
one (or more) of the following papers in your work :
[1] L. Noe, G. Kucherov,
YASS: enhancing the sensitivity of DNA similarity search,
2005, Nucleic Acids Research, 33(2):W540-W543.
[2] L. Noe, G. Kucherov,
Improved hit criteria for DNA local alignment,
2004, BMC Bioinformatics, 5:149.
