-
Notifications
You must be signed in to change notification settings - Fork 16
/
NEWS.TXT
29 lines (28 loc) · 1.83 KB
/
NEWS.TXT
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
################################# PIRANHA NEWS #################################
New in Piranha v1.2.1
* This news file was added to keep track of updates to Piranha
* Piranha now preserves the strand by default. If user wishes to output all
the peaks on positive strand, the -x option will do so.
New in Piranha v1.2.0
* Piranha now outputs only the significant sites as per a threshold on the
p-value of a site. By default that threshold is 0.01, but it can be set
using the new option -p. To make piranha output all of the sites (i.e.
make Piranha behave in the same way as earlier versions), use -p 1.
* All p-values output by Piranha are now automatically corrected for
multiple hypothesis testing using the method of Benjamini and Hochberg.
This can be turned off using the -c option, which will make Piranha
output uncorrected p-values.
* By default, only the smallest 99% of positions are now used for fitting;
they are considered to represent the background. This value was empirically
arrived at by analysis of a collection of CLIP-Seq datasets. However,
it can be manually set using the new -a option. To use all of the input
data for fitting, you would use "-a 1" to set the threshold. If you find
that too few locations are significant, you may try decreasing this value.
* It is now possible to use piranha to bin covariates as well as the
response file (although the default is still to assume both response and
covariate files are already binned). When binning covariates (i.e.
assuming they are raw reads), Piranha will automatically populate missing
bins in the covariate file.
* Piranha now automatically merges adjacent significant bins
* By default, Piranha now outputs covariate values with each bin, but
this can be disabled with the -r option