Skip to main content

Table 5 Assessment of running times for all tools. Resource consumptions for LazyB are shown for the complete assembly process comprising (1) ABySS unitig assembly; (2) Mapping of unitigs to long reads and (3) running LazyB itself, denoted by A+m+ LazyB, and the the last step only, denoted by LazyB

From: LazyB: fast and cheap genome assembly

\(\times \)

Tool

Time

RAM (MB)

Yeast

 

ABySS unitig

00:00:11:03

2283

 

ABySS full

00:00:20:01

2283

\({\sim} 5\times \)

Mapping

00:00:00:05

540

LazyB

00:00:00:30

136

A+m+ LazyB

00:00:11:38

2283

Canu

00:10:23:55

2617

Wtdbg2

00:00:13:08

698

HASLR

00:00:06:44

4922

DBG2OL

00:00:31:46

1141

Wengan

00:00:06:45

4400

A+Wengan

00:00:26:46

4400

\({\sim} 11\times \)

Mapping

00:00:00:15

1544

LazyB

00:00:01:46

362

A+m+ LazyB

00:00:13:04

2283

Canu

00:13:44:16

6779

Wtdbg2

00:00:29:28

1142

HASLR

00:00:08:09

4922

DBG2OL

00:00:51:13

1264

Wengan

00:00:14:29

4421

A+Wengan

00:00:34:30

4421

Fruit fly

 

ABySS untig

00:02:32:39

25344

 

ABySS full

00:04:56:03

25346

\({\sim} 5\times \)

Mapping

00:00:02:43

6433

LazyB

00:00:08:33

613

A+m+ LazyB

00:02:43:55

25344

Canu

02:13:51:39

7531

Wtdbg2

00:01:37:41

3395

HASLR

00:01:30:33

5531

DBG2OL

00:07:58:22

6151

Wengan

00:01:41:26

5394

A+Wengan

00:06:10:29

25346

\({\sim} 10\times \)

Mapping

00:00:06:11

9491

LazyB

00:00:11:57

2241

A+m+ LazyB

00:02:50:47

25344

Canu

07:04:08:28

7541

Wtdbg2

00:04:02:43

5024

HASLR

00:01:43:21

5553

DBG2OL

02:07:32:01

17171

Wengan

00:02:28:51

5323

A+Wengan

00:07:24:54

25346

  1. Step (1) is often not needed as short-read assemblies are available for many organisms
  2. Similarly, Wengan requires a full ABySS assembly as a its basis. Resources are only compared for yeast and fruit fly, because Canu cannot be run for human in acceptable time and resource-constraints on our equipment. As all tools except LazyB and DBG2OL are parallelized, running times are given as the sum of time spent by all CPUs. Therefore, computational effort is measured rather than wallclock time. ABySS greatly dominates the LazyB pipeline and to a lesser degree also Wengan. Nevertheless, LazyB is faster by a factor of \(>60\) compared to Canu, \(\approx 3\) compared to DBG2OL, and \(\approx 2.5\) to Wengan