Commits

Author Commit Message Labels Comments Date
chuttenh
[svn r369] Fix COALESCE cutoffs to prefer two end gaps to a mismatch Add exponential downweighting of partial PST matches Fix motif counting in CCoalesce* to include PST motifs
chuttenh
[svn r368] Add kmer/RC to PST merging to CCoalesceMotifLibraryImpl Also added smarter string/PST and PST/PST alignment for this in CPST Fix (maybe?) the motif p-value multiple hypothesis correction Fix a ridiculous floating point rounding bug in CCoalesceHistogramSet::CohensD Improve PST matching scores at sequence boundaries Add motif count threshhold for merging to COALESCE This vastly improves performance for clusters with lots of motifs
chuttenh
[svn r367] Improve sequence alignment and RC/PST management in CCoalesceMotifLibrary Add RC/RC and kmer/RC merging to COALESCE Fix an OBO in CCoalesceHistogramSet::Add
chuttenh
[svn r366] Add rudimentary motif merging to PSTs in COALESCE Currently only kmer/kmer merges happen! To merge, must be distributed identically below some p-value Must also align below some edit distance cutoff Add minimum cluster size to COALESCE Clusters below minimum size will be processed but not finalized Motif detection is not performed while processing clusters below minimum Add column normalization to CPCL (per-condition z-score) Add defaul…
chuttenh
[svn r365] Add large cluster balancing to CCoalesce Add a rudimentary binary file format to CPCL Add sequence type tracking to CFASTA Add exponential PDF and some documentation tweaks to CStatistics Improve behavior for zero conditions or missing sequence in CCoalesce
chuttenh
[svn r364] Improve CStatistics sum/sum squared/average/variance organization Improve histogram sum/sum squared/average/variance organization Add z-score weighting and average sequence subtraction to COALESCE
chuttenh
[svn r363] Add motif "subtraction" to COALESCE Remove randomized bootstrapping from COALESCE All it was doing was an extra multiple hypothesis correction Much faster to do that directly
chuttenh
[svn r362] Add semi-functional randomziation based p-value adjustment to COALESCE It does something, possibly something good, but not something perfect Mostly a safekeeping checkin in order to try more drastic measures safely
chuttenh
[svn r361] Safekeeping checkin before attempting various motif management changes Add reverse complement generation and tracking to CCoalesceMotifLibrary This could be done more elegantly, but I haven't been able to figure out how
chuttenh
[svn r360] Fix a bug in the COALESCE histogram ave/std calculation
chuttenh
[svn r359] Add very simple sequence cache to speed repeated startups of COALESCE Fix COALESCE motif histograms for genes with no available sequence info Fix some stats tweaks in COALESCE Back to Poisson/T-test for motifs; Chi2 too sensitive for large sample sizes Fix something that I completely screwed up when calculating P(g in C|S)
chuttenh
[svn r358] Improve Sleipnir COALESCE performance Uses ~1/3 more memory to improve precision But about 10-100x faster than my original port Fix histograms for sequences of different lengths Binned in configurable increments of hits/base units Improve Bayesian integration of sequence and expression data Full Bayesian with independence assumption, no prior Add multiple hypothesis testing and make chi2 test default in COALESCE This seems t…
chuttenh
[svn r357] Fix Linux burps with COALESCE code
chuttenh
[svn r356] Fix a dumb crashing bug in MIer (missing null pointer check) Fix a minor bug in BNs2Txt (nodes not IDed correctly from given datasets)
chuttenh
[svn r355] Add incremental histogram updates for big speed improvements
chuttenh
[svn r354] Improve per-gene kmer count management Reduces memory usage, improves speed Safekeeping checkin before seeing if this is better incrementally
chuttenh
[svn r353] Add motif/sequence tracking to COALESCE Tested to work ok, no speed optimization yet Add CCoalesceMotifLibrary to handle motif discovery Only manages kmer<->integer mapping now Set up to do merging, non-kmer motifs, and known motif naming Improve COALESCE intermediate and debug output management Add ModifiedBessel and Skellam functions to statistics Fix Kolmogorov-Smirnov test index bug and p-value weirdness NR disagrees with R;…
chuttenh
[svn r352] Fix some Linux quibbles with the COALESCE additions
chuttenh
[svn r351] Add preliminary COALESCE implementation Expression clustering appears to work, minor speed optimization done Sequence clustering untested Add CFASTA for virtualized FASTA file parsing Tested reasonably well Aware of different sequence types (3'/5'/ORF) and subtypes (introns/exons) Improve KS test p-value computation based on Coalesce code Also fix a dumb sort that was hurting efficiency Add second gene set lookup to Dat2Dab Ad…
chuttenh
[svn r350] Update to latest version of SMILE
chuttenh
[svn r349] Add 1.1 version history info
chuttenh
[svn r348] Add BNs2Txt and Mat2Txt documentation Add CDatFilter and new stats documentation
chuttenh
[svn r346] Fix Linux cmdline.[ch]s in preparation for 1.1 tag
chuttenh
[svn r345] Fix GO parser to play nice with new versions and not complain so much
chuttenh
[svn r344] Add term-internal gene counts to DChecker Add random subsampling of output to Dat2Dab Add full Winsorization to ratio gene set queries in BNServer
chuttenh
[svn r343] Add Linux build updates for BNs2Txt Fix a few typos
chuttenh
[svn r342] Add configurable normalization to CDat Update Counter to new contexts.txt structure Add PMID to Sleipnir Add single set within/background calculation to Funcaeologist Fix borked BNServer within and between calculations for custom gene sets Add BNs2Txt tool for dumping binary minimal Bayes nets to (X)DSL files
chuttenh
[svn r341] Fix various Linux glitches with the latest checkin
chuttenh
[svn r340] Add a gene exclusion filter to Dat2Dab Add bootstrap null distribution calculations to Funcaeologist Randomly picks subgraphs of requested sizes Calculates stdev of between / background / within * prior score Add between/within/background calculations to Hubber Precalculates scores for or between given gene sets or genomic background Add Mat2Txt tool for dumping saved binary matrices Add percentile, Winsorization, and Welch's t-t…
chuttenh
[svn r339] Improve BNServer organization prior to incorporating p-values Standardize GetAssociation and in/out caluations Consolidate several utility methods Slightly improve memory management and gene indexing Add pair counting across datasets to Dab2Dad Improve background calculation in Funcaeologist Out denominator is now all genes in either set Improve efficiency and locking in CDatabaselet Add pair counting above threshhold in Dat2Dab Fix stupid table outp…
  1. Prev
  2. Next