UCHIME improves sensitivity and speed of chimera detection. Robert C. Edgar and colleagues developed UCHIME, a new program for detecting chimeric sequences in DNA samples. Chimeric sequences often form during PCR amplification, especially when sequencing single regions like 16S rRNA or fungal ITS. Undetected chimeras can be misinterpreted as novel species, leading to inaccurate diversity estimates. UCHIME uses a database of non-chimeric sequences or detects chimeras de novo by analyzing abundance data. It outperforms existing methods like ChimeraSlayer and Perseus in sensitivity, especially with short, noisy sequences. UCHIME is significantly faster than these methods, with execution times 100× faster than Perseus and 1000× faster than ChimeraSlayer. UCHIME can use a reference database or operate in de novo mode, making it versatile for different sequence types. It reports scores for each sequence, allowing users to adjust sensitivity and specificity by changing the minimum score threshold. UCHIME does not require training, as its parameters are robust across different input data. The default score threshold provides good sensitivity with low error rates. UCHIME's algorithm divides the query sequence into four segments, searches a reference database, and constructs a three-way alignment to detect chimeras. It identifies three types of chimeric alignments: local, local-X, and global-X. UCHIME uses a scoring function based on differences between the query and candidate parents, with a threshold to determine if a sequence is chimeric. It also selects candidate parents based on abundance, assuming chimeras are less abundant than their parents. UCHIME was tested on various datasets, including SIM2, SIMM, and MOCK, showing improved sensitivity and accuracy compared to existing methods. UCHIME is efficient and robust, making it a valuable tool for detecting chimeras in sequence analysis.UCHIME improves sensitivity and speed of chimera detection. Robert C. Edgar and colleagues developed UCHIME, a new program for detecting chimeric sequences in DNA samples. Chimeric sequences often form during PCR amplification, especially when sequencing single regions like 16S rRNA or fungal ITS. Undetected chimeras can be misinterpreted as novel species, leading to inaccurate diversity estimates. UCHIME uses a database of non-chimeric sequences or detects chimeras de novo by analyzing abundance data. It outperforms existing methods like ChimeraSlayer and Perseus in sensitivity, especially with short, noisy sequences. UCHIME is significantly faster than these methods, with execution times 100× faster than Perseus and 1000× faster than ChimeraSlayer. UCHIME can use a reference database or operate in de novo mode, making it versatile for different sequence types. It reports scores for each sequence, allowing users to adjust sensitivity and specificity by changing the minimum score threshold. UCHIME does not require training, as its parameters are robust across different input data. The default score threshold provides good sensitivity with low error rates. UCHIME's algorithm divides the query sequence into four segments, searches a reference database, and constructs a three-way alignment to detect chimeras. It identifies three types of chimeric alignments: local, local-X, and global-X. UCHIME uses a scoring function based on differences between the query and candidate parents, with a threshold to determine if a sequence is chimeric. It also selects candidate parents based on abundance, assuming chimeras are less abundant than their parents. UCHIME was tested on various datasets, including SIM2, SIMM, and MOCK, showing improved sensitivity and accuracy compared to existing methods. UCHIME is efficient and robust, making it a valuable tool for detecting chimeras in sequence analysis.