TCRcloud 1.3.1

TCRcloud is a TCR repertoire visualization and comparison tool

TCRcloud is a TCR repertoire visualization and comparison tool

Instalation

TCRcloud is written in python and can be installed from PyPI using pip:

pip3 install TCRcloud

Currently it is only compatible with Linux (x86-64) and macOS (x86-64) because one of the dependencies used is also only compatible with those operating systems. It does work on WSL if you want to use Windows.

TCRcloud uses the AIRR Data Commons API and needs AIRR compliant data as input.

TCRcloud was initially developed for TCR repertoires but it is also compatible with BCR repertoires.

Usage

To create a word cloud

TCRcloud cloud repertoire.airr.rearrangements.tsv

By default TCRcloud colours the CDR3 based on the Vgene but you can provide a json file that atributes colours in Hex format to specific sequences:

{
"#FF0000":["CAASITGNQFYF","CAVREDGTSGSARQLTF"],
"#0000FF":["CAVMDSNYQLIW"]
}

The sequences not in the file will be coloured grey.

Only the colours for human TCR and BCR variable genes are coded into TCRcloud.

To use your custom colours for the word cloud

TCRcloud cloud repertoire.airr.rearrangements.tsv -c colours.json

To create a word cloud without a legend

TCRcloud cloud repertoire.airr.rearrangements.tsv -l False

To create a radar plot comparing diversity metrics

TCRcloud radar repertoire.airr.rearrangements.tsv

By default TCRcloud uses repertoire_id but you can create a legend with the text you want by providing a json file:

{
"2839362682105696746-242ac113-0001-012":"Twin 2A",
"2939134772391776746-242ac113-0001-012":"Twin 2B"
}  

To create a radar plot with your desired legend

TCRcloud radar repertoire.airr.rearrangements.tsv -c legend.json

To create a radar plot without a legend

TCRcloud radar repertoire.airr.rearrangements.tsv -l False

To export the calculated metrics from the radar to a text file

TCRcloud radar repertoire.airr.rearrangements.tsv -e True

The metrics utilized in the radar

Distinct CDR3: A simple count of the unique CDR3 sequences in the sample

D50 Index: Developed by Dr Jian Han, this metric represents the percent of dominant and unique T or B cell clones that account for the cumulative 50% of the total CDR3 counted in the sample

Convergence: Frequency of CDR3 amino acid sequences that are coded by more than one nucleotide sequence

Gini Index: Originally developed by Dr Corrado Gini to represent the wealth inequality within a social group, this metric is a measure of distribution, with 0 representing perfect equality and 1 representing perfect inequality between CDR3

Shannon Index: Originally developed by Dr Claude Shannon to quantify the entropy in strings of text, this metric takes into account the number of CDR3 present, as well as, the relative abundance of each CDR3 and higher the number, the higher is the species diversity

Simpson Index: Originally developed by Dr Edward H. Simpson to measure diversity in a ecosystem, this metric measures the probability that two randomly selected CDR3 are different

Chao1 index: Originally developed by Dr Anne Chao to estimate richness in a community, this metric indicates the estimated number of CDR3 in a sample

Using TCRcloud you can download rearragements files from the AIRR compliant databases based on AIRR repertoire metadata files

To download AIRR rearrangements files

TCRcloud download repertoire.airr.json

TCRcloud provides some test data to experiment the tool. The data is one twin pair from the monozygotic twins study from the Mark Davis lab (DOI: 10.1038/ncomms11112)

To download the test data repertoire file

TCRcloud testdata

After having the testdata.airr.json file you can use the download function included in TCRcloud to get the matching rearragements file.

Examples:

TRA CDR3 word cloud

TRB CDR3 word cloud

TRG CDR3 word cloud

TRD CDR3 word cloud

Diversity comparison

Comparing here the αβ repertoire from one twin pair from the monozygotic twins study from the Mark Davis lab (DOI:10.1038/ncomms11112)