Frequently Asked Questions

What kind of data can I use Lambda with?

Lambda has three modes:

1. protein mode: all searches in protein databases (Uniprot, Uniref, …). Query sequences can be protein or DNA/RNA (WGS, WES, …). This covers the Blast programs BlastP, BlastX, TBlastN and TBlastX.
2. nucleotide mode: searches of regular DNA/RNA sequences in DNA databases. This covers the Blast programs BlastN and Megablast.
3. bisulfite mode: searches of bisulfite treated reads in DNA databases.

What types of read data are supported?

Lambda is tested mostly on Illumina short read data. The algorithm is currently not optimised for long-read data.

What kind of speed-ups can I expect over NCBI Blast?

This depends on the dataset, the number of query sequences and the size of the database. We measured speed-ups of over 200x with the default profile and over 2000x with Lambda's fast profile.

What kind of sensitivity can I expect compared NCBI Blast?

The default profile should achieve 95-99% of the sensitivity, but this also depends on highly on the dataset. We recommend testing with a small sub-sample if the sensitive is absolutely crucial to you.

What are the system requirement of Lambda?

The more cores you have, the better! Memory requirements are about $(qSize + dbSize) * 1.2$ where $qSize$ is the the size of the uncompressed query file and $dbSize$ is the size of the uncompressed index file.

The indexer has additional memory requirements.