Skip to main content

Table 3 Performance of the LAST+MEGAN-LR pipeline

From: MEGAN-LR: new algorithms allow accurate binning and easy interactive exploration of metagenomic long reads and contigs

Step

Input

Output

Runtime

Memory

PacBio reads on HMP mock community

Align

Reads file (1.5 GB)

MAF file

119 min

23 GB

Convert

MAF file (49 GB)

DAA file

29 min

5 GB

Classify

DAA file (4.2 GB)

Meganized DAA file (4.5 GB)

6 min

5 GB

PacBio reads on Singer et al. mock community

Align

Reads file (56 MB)

MAF file

10 min

22 GB

Convert

MAF file (8.9 GB)

DAA file

5 min

5 GB

Classify

DAA file (197 MB)

Meganized DAA file (415 MB)

1 min

5 GB

Nanopore reads on HMP mock community

Align

Reads file (191 MB)

MAF file

10 min

22 GB

Convert

MAF file (6.1 GB)

DAA file

3 min

5 GB

Classify

DAA file (553 MB)

Meganized DAA file (644 MB)

1 min

5 GB

Anammox data

Align

Reads file (336 MB)

MAF file

31 min

24 GB

Convert

MAF file (8.5 GB)

DAA file

4 min

5 GB

Classify

DAA file (371 MB)

Meganized DAA file (500 MB)

2 min

5 GB

Cheese data

Align

Reads file (5.1 GB)

MAF file

251 min

24 GB

Convert

MAF file (93 GB)

DAA file

90 min

10 GB

Classify

DAA file (3.1 GB)

Meganized DAA file (3.5 GB)

5 min

10 GB

  1. For each of five long read datasets, we report the wall-clock time and main memory required by LAST to align against the NCBI-nr database, for MEGAN to convert the LAST MAF output files into DAA format, and then for MEGAN to classify the reads so as to meganize the DAA file, respectively. The computations were performed on a Linux server with 32 cores and 512GB memory