TABLE 1.
Characteristics | Genefinder | Mykrobe (9) | Typewriter (10) |
---|---|---|---|
Method | Maps raw reads to list of target alleles using Bowtie | Looks for list of target alleles in de Bruijn assembly graph | Blasts list of target alleles against de novo assembliesa |
Input | Fastq file | Fastq file | Genome assembly output (Velvet) |
Required homology to declare gene presence/absence | >90% to target allele | Based on kmer recovery: K is minimum percentage expected to be recovered for a gene; K = 0.3 for blaZ, K = 0.6 for fusB and fusC, K = 0.8 otherwise | >90% relative coverage (homology by length) (80% for blaZ) |
Required homology to declare SNP | >90% to target: can be modified | 100% of 63-kmers required to call a variant present | >90% to target: can be modified |
Prediction of stop codons in genes present | Yes | No: there is no assembly | Yes |
Reads can be mapped to | Multiple targets | Single target | Single target |
Speed/processor | 1 to 3 min on laptop with 2.3-GHz processor and 16-GB memoryb | 2 min on laptop with 2.3-GHz processor and 16-GB memory | 3 h for assemblies on cloud computational system, then a few minutes for BLAST |
Sequence quality control | Threshold adjusted if gene has multiple reference sequences or variable level of diversity; can detect potential contamination by comparing avg depth of coverage | Can identify mixtures of different species and same species | Thresholds for N50 and parallel reference-based mapping: nothing reported if below these thresholds |
Using blastn for sequence identity and tblast for mutations.
Genefinder speed is relative to the number of genes present in the database.