ReleaseHistory

v1.6.15 first for public release

v1.7.01 improved the chunk reader; When identifying signatures of insertions, PoPooaltionTE2 reads chunks of the ppileup in order to avoid excessive memory consumption. This version has an improved algorithm which ensures that a true separate analysis of files results in the same results as --mode separate

v1.8.02 Two new options; allows to sample the ppileup tracks with replacment; Very cool new option: --homogenize-pairs for generating the ppileup track; This option allows to use the same number of paired-ends for each sample. Thus if someone wants to use the same number of reads in all samples, it is not necessary anymore to perform time-consuming counting of mapped pairs in all bam-files (better would be informative pairs) and than to randomly sample the same number of pairs from the bam file (Note sampling of pairs from a bam is quite painful).

v1.9.03 New algorithm for pairing up signatures of TE insertions; much faster than before; minor bug fix with options of identify signatures

v1.10.03 Eliminated a memory leak during generating the ppileup. Generating ppileup now requires only a minimum amount of memory, like 1GB RAM for 8 populations in a joint analysis. Moreover this eliminated a small bug during ppileup generation. The population frequency is now more accurately estimated. The accuracy of the position estimate, true positives and false positives is not affected. Many thanks to Emmanuelle Lerat for pointing the problem out.