VSEARCH 1.4.7

Fixes a problem with searching for or clustering sequences with repeats. In this new version, vsearch will look at all words occurring at least once in the sequences in the initial step. Previously only words occurring exactly once were considered. In addition, vsearch now requires at least 10 words to be shared by the sequences, previously only 6 were required. If the query contains less than 10 words, all words must be present for a match. This change seems to lead to slightly reduced recall, but somewhat increased precision, ending up with slightly improved overall accuracy.

VSEARCH 1.4.6

Fix bug in SIMD aligner introduced in VSEARCH version 1.2.7 leading to reduced accuracy

VSEARCH 1.4.5

Fix memory allocation bug when reading long FASTA sequences

VSEARCH 1.4.4

Remove debug message

VSEARCH 1.4.3

Fix a bug with determining amount of memory on some versions of Apple OS X.

VSEARCH 1.4.2

Dynamic loading of compression libraries

VSEARCH 1.4.1

Added public domain MD5 and SHA1 code to eliminate dependency on crypto and openssl libraries and their licensing issues.

VSEARCH 1.4.0

Several new commands and option added, including subsampling and fastq file processing. New autotools build system.

VSEARCH version 1.1.3

More bug fixes for large datasets

VSEARCH version 1.1.2

Fix bug with all commands involving searches in large datasets (>4Gbp). Fix format of help text.