v1.17.9

• removed g++ v4.7 requirement and hard-coded paths to GSL library in Makefile
• included a simulated dataset

v1.17.8

• added checks whether files are open for writing
• changed to new defaults: --(cna/baf/snv)-rnd [double:1.0e-6] (nan)
• allowed --cna-jump -1 and --baf-jump -1 (no jumps)
• --cna-jumps [baf-jumps-file] and vice versa enabled (useful for exome data)
• jumps read and integrated with new function match_jumps() (not get_track()).
• fixed bug when chromosomes have no non-zero observations.

v1.17.7

• fixed range error in pre-filter in pick-from/match-to mode.

v1.17.6

• fixed nan bug in GOF, when N==0 (missing data).
• fixed bugs in pre-filter, when --window-size is greater than length
• fixed bug in pre-filter in pick-from-match-to mode

v1.17.5

• fixed memory alloc bug in pre-filter
• abandon ftp site for releases, used only for backup and beta

v1.17.4

• fixed fatal bug in snv-mode with correlations

v1.17.3

• new program pre-filter
• --snv-pen to --snv-pen-high and --snv-pen-mult
• --baf-pen to --baf-pen-compl
• --cna-pen to --cna-pen-zero, --cna-pen-diff and cna-pen-norm
• split README
• fixed bug in SNV transition matrix in combination with --max-tcn [file]

v1.17.2

• new output: posterior per subclone, goodness of fit (GOF) per segment
• changed file name *clonal.txt -> *summary.txt
• filterHD STDOUT includes now GOF per sample
• cloneHD *summary.txt includes now GOF per sample
• changed _ to - in all file names
• fixed bug: BAF now symmetrized only in per-subclone-posterior
• new CNA prior to penalize homozygous deletions --cna-pen [double:0.9]

v1.17.1

• BAF posterior symmetrized for output
• CNA transition matrix penalizes clones with zero copies of a segment
• fixed bug in SNV prior computation
• added pre-processor directives for conditional openMP compilation

v1.17.0

changed the way SNV priors are computed:

• if CNA given: SNV prior informed by CNA posterior
• if CNA+BAF given, SNV prior informed by BAF+CNA posterior
• if SNV only and --max-tcn not given, assumes all chr to be all-normal, mean total c.n. to be normal; SNV prior parameters can be learned with --learn-priors 1.
• if SNV only and --max-tcn [int/file] is given, this data is used to fix the total c.n. per chr and subclone; mean total c.n. is calculated on the fly; SNV prior parameters can be learned with --learn-priors 1.
• if SNV only and --max-tcn [int/file] and --avail-cn [file] are given, SNV prior is calculated according to c.n. availability.

more changes

• changed option --copynumber [file] to --mean-tcn [file]
• new option --avail-cn [file]
• changed option --maxcn [int:4] to --max-tcn [file/int]
• changed option --snv-err [double] to --snv-fpfreq [double]
• changed option --snv-fpr [double] to --snv-fprate [double]
• output file *used-tcn.txt to *used_mean_tcn.txt
• output file *copynumber.txt to *mean_tcn.txt
• new output file *available_cn.txt
• changed sample to chr in cloneHD output files
• slimmed down output of --print-options.
• split clone.cpp into components clone-*.cpp
• split off cloneHD-inference.cpp
• new Makefile