What's new in CoNet 1.0 Beta 6
Sep 1, 2014
- change of license (from 22nd May 2014 onwards, CoNet is distributed under the GNU General Public License version 2)
- fixed bug in Kendall implementation
- treatment of ties in Spearman implementation fixed
- fixed problem of setting initial edge selection parameter from settings file
- if groups are provided, the group properties are listed when matrix info is selected
- improved taxonomy parsing from biom files (hit number is parsed if provided)
- fixed rarefaction bug (previously, error message prevented to carry out rarefaction)
- network comment does no longer list removed edges and filtered rows, to reduce size of big networks
- logging level can now be adjusted (defaults to fatal to reduce size of log file)
New in CoNet 1.0 Beta 5 (May 5, 2014)
- added p-value computation with Fisher's Z-score for correlations
- Lallich routine selection without multiple test correction rejected - bug fixed
- parsing error for sample names starting with E/e and followed by numeric value fixed
- fixed positive-edges-only filter
- fixed computation of maximum number of zeros per row (for matrix property display)
- added error message in case automatically obtained lower and upper threshold overlap
New in CoNet 1.0 Beta 4 (May 5, 2014)
- implemented export of positive and negative node degree
- option added to report input matrix properties and suggestion of CoNet settings based on these properties
- fixed settings file bug (disable speedup not set/exported from/to settings file)
- fixed settings file bug (minzeropairs filter not set from settings file)
- fixed settings file bug (copresence/mutual exclusion-only-filter not set back correctly to default)
- fixed command line export bug (--topbottom not exported)
- fixed treatment of feature rows for taxa without lineages in renormalization
- edges for which methods infer conflicting interaction types are filtered now
- row combination involving summed row are forbidden now
New in CoNet 1.0 Beta 3.1 (Nov 19, 2013)
- Jean-Sebastien Lerat's library KLD implementation: removed division by number of value pairs
- bug fix for networks merged on the score level
- improved non-SGE parallelization support
New in CoNet 1.0 Beta 3 (Nov 19, 2013)
- fixed bug concerning automatic lineage assignment for higher-level taxa with groups
- allow transposing feature matrix
- allow feature/input matrix sample matching to support feature data with samples that have a different order than input matrix samples
- more flexible OTU table from biom file parsing (support parsing lineages from row names)
- better matrix parsing error catching
- too large data menu size fixed
- prevent generation of command line call in the presence of input errors
- removed faulty error message about combining ARACNE with missing values when JSL MI was selected
New in CoNet 1.0 Beta 2 (Nov 19, 2013)
- support for MI and Jensen-Shannon with Jean-Sebastien Lerat's library
- interface offers 3 MI implementations: minet, aracne and Jean-Sebastien Lerat's library
- fixed bug related to keeping discarded rows as one row per group
- fixed bug concerning group exclusion (betweengroupsonly, withingroupsonly)
- prevent kept summed rows to appear in the network
New in CoNet 1.0 Beta 1.1 (Nov 19, 2013)
- No Blast hit entry supported in QIIME OTU tables
- execution of R batch scripts in given directory
New in CoNet 1.0 Beta 1 (Nov 19, 2013)
- This release features the re-implementation of the CoNet core by Jean-Sebastien Lerat.
- The re-implementation included the renormalization and the measures:
- Pearson
- Spearman
- Kendall
- Hellinger
- Euclid
- Steinhaus
- Bray Curtis
- Kullback-Leibler
- Variance of Log-ratios
- Hilbert-Schmidt independence criterion
- All measures were tested to give the same networks from test data as the previous implementations.
- This release also adds the following features:
- support of QIIME OTU tables with and without taxonomy
- versioning via CoNet property file, valid for both Cytoscape plugin and command line version
- added the possibility to filter rows with less than the given number of non-double-zero pairs
- directory of R batch scripts can be specified on command line