Datasets and results to evaluate methods for analysing immunoglobulin sequences
HTML Jupyter Notebook
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is even with sdwfrost:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
data
results
LICENSE
README.md

README.md

igh-evaluation

Datasets and results to evaluate methods for analysing immunoglobulin sequences

Data

  • aborigin_simulations: A random selection of 1000 simulated sequences used to evaluate Ab-origin, downloaded from here.

  • igblast_vmatch: These data were used by Ye et al. (2013) to evaluate IgBLAST, and comprise of 100 IGH sequences with no mutations in IGHV, hence these sequences should contain few or no mutations in the D and J regions. These data were downloaded as a PDF from the supplementary information here and copied to a text file.

  • igscueal_simulations: These are simulated human IGH sequences described by Frost et al. (2015).

  • N152: This is a dataset of 11 clonally related IGH sequences taken from an HIV-1 infected donor N152, from whom the broadly neutralizing antibody 10E8 was isolated (Zhu et al (2013)). The sequences were downloaded from GenBank.

  • ohmlaursen: This is a dataset of 6329 clonally unrelated IGH sequences obtained from individuals homozygous for IGHV3-23*01 and IGHJ6*02, and amplified using primers that were intended to be specific for IGHV3-23. The original study also amplified a number of IGHV3-h pseudogenes, which have been excluded from the file. This dataset was used to evaluate Ab-Origin. The full dataset (including pseudogenes) was obtained from GenBank.

  • PNG: 1108 human IGH sequences from individuals in Papua New Guinea, known to contain clonally related sequences; a CSV file containing germline assignments was downloaded from here. Sequences were downloaded from GenBank directly, and sequences and annotations extracted using an IJulia notebook. Unfortunately, the (manual) assignments of sequences to clones is not available.

  • PW57: A dataset of 57 clonally related IGH sequences from IgD+ IgM-CD38+ B cells described by Wilson et al. (2000), corresponding to GenBank accessions AF262145–AF262201, extracted from PopSet 8810007.

  • PW99: A dataset of 106 clonally related IGH sequences from IGD+ IGM- CD38+ B cells described by Zheng et al. (2005), corresponding to GenBank accessions EF544883-EF544988. This dataset is often referred to as PW99; however, I was unable to determine which sequences were excluded from this larger set.

  • S22: The Stanford S22 dataset is a set of 13,153 human IGH sequences derived from an individual who was fully genotyped. The performance of a utility is determined by the proportion of sequences that are assigned to a germline gene that is absent from the individual. These files were downloaded from the iHMMune website.