Performance

All runs were performed using Trinity 2.1.1 with 120G of RAM, 16 CPU's, and the --bflyCalculateCPU parameter. Each data set consists of approximately 50 million reads taken from the described organism. Tests were performed on Indiana University's Karst cluster, which consists of IBM NeXtScale nx360 M4 servers, each equipped with two Intel Xeon E5-2650 v2 8-core processors. Tests were performed on nodes with 512GB of memory.

Transcript Reconstruction Performance

Since Trinity attempts to reconstruct RNA transcripts, not a genome, there is no single sequence for what Trinity calls a gene. A Trinity gene is a collection of related transcripts. In a de novo assembly, the transcripts are related by having shared kmers during the assembly stage.

Runtime

The total amount of wall time required to produce an assembly. S. pombe times do not include jaccard clipping.

The many components of Trinity each have their own unique processor requirements. Additional information can be found on the Trinity Processor Usage page.

Memory Usage

The maximum amount of memory used at a single time when producing an assembly. This number does not include memory used by the parallel Butterfly component, which can use as much or as little memory as is available on the system. (See Trinity Memory Usage for more information)

The many components of Trinity each have their own unique memory requirements. Additional information can be found on the Trinity Memory Usage page.

References

1. Bo Li, Nathanael Fillmore, Yongsheng Bai, Mike Collins, James A. Thompson, Ron Stewart, Colin N. Dewey. Evaluation of de novo transcriptome assemblies from RNA-Seq data.