Trimadap is a small tool to trim adapter sequences from Illumina data. It performs SSE2-SW between each read and each adapter sequence and identifies adapter sequences with a few heuristic rules which can be found in the ta_trim1() function in trimadap-mt.c. The default adapters it uses are included in illumina.txt. These are typical Illumina adapters from paired-end sequencing.

Trimadap is designed as an on-the-fly stream filter. It is very fast. In the multi-threading mode, it is as fast as reading through a gzip-compressed FASTQ file. On the other hand, trimadap is very conservative. It is not good in terms of accuracy as of now. I will probably fine tune the heuristic rules in future. This should not be hard in principle, but it takes development time.