| 11 genomes | 30 genomes | 100 genomes |
---|
 | Original genomes | Filtered genomes | Original genomes | Filtered genomes | Original genomes | Filtered genomes |
FS = 500 bps | 71.10% | 77.30% | 51. 6% | 55.70% | 40.50% | 41.10% |
FS = 1000 bps | 79.90% | 85.90% | 65.30% | 70.30% | 51.10% | 52.60% |
FS = 2000 bps | 86.30% | 91.70% | 74.80% | 80.60% | 61.00% | 68.53% |
FS = 5000 bps | 91.10% | 98.10% | 86.60% | 93.20% | 79.40% | 81.90% |
FS = 10000 bps | 95.80% | 99.30% | 91.90% | 97.50% | 86.60% | 89.18% |
- The binning accuracy is defined as (prediction specificity + prediction sensitivity)/2, and FS is for f ragment s ize, where both the specificity and sensitivity are measured in terms of putting the fragments into the correct bin corresponding to each genome, defined by the majority of the fragments in the bin. The column "Original genomes" lists the binning accuracy of our algorithm on all the non-overlapping fragments in each group of genomes, and the column "Filtered genomes" gives the accuracy after removing the 10% fragments with the most abnormal barcodes from each genome.