TranslocatoR

TranslocatoR finds translocations in MosaiCatcher-processed data. It can be used for both reciprocal and non-reciprocal translocations.

In this README:

• How TranslocatoR Works
• Getting Started
• Using TranslocatoR
• Ready-to-use Example

How TranslocatoR Works

After processing the data with StrandPhaseR, the MosaiCatcher pipeline outputs a file containing all segments and their phased states. This file forms the input to TranslocatoR.
TranslocatoR has several ways for the detection of co-segregating segments. Besides the two main options, pq and majority, a further file of additional inferred strand-states can be provided by the user (trfile) as well as a file containing regions of interest in the sample.
TranslocatoR performs haplotype splitting for each segment: Note that with the majority option, each chromosome is represented by one ‘segment’ which is the majority state, and with the pq option each chromosome is represented by two segments, one for each arm. Finally each haplotype of each segment is compared to every other segment’s haplotype.

TranslocatoR outputs a matrix containing translocations by selecting all segments where p is below a user-chosen cutoff (default 0.01).

graph LR

n1(Strand-Seq) -->|BAM files| n2[MosaiCatcher]
n2 --> |Strand States| n4{TranslocatoR}
n3(manual strand state identification)-.trfile.->  n4
n4 -->n5(translocation)
n6(regions of interest)-.regions.->  n4

Loading

Getting Started

Installing from GitHub

If you want to use TranslocatoR directly without looking at the usage example first, you can install it directly from GitHub with the devtools() package. Simply run devtools::install_github(repo = "https://github.com/Sanders-Lab/TranslocatoR", ref = "master"), load the TranslocatoR library with library(TranslocatoR) and you're ready to go. Have fun!

If you would like to do an example run first to get a feel for how TranslocatoR output looks, you need to clone the directory and install the package separately. I have not yet been able to figure out how to make devtools install the subfolders necessary for the example to run. Clone the whole repository into your desired local directory by navigating to it, then running git clone --recurse-submodules https://github.com/Sanders-Lab/TranslocatoR.git. Inside Terminal or RStudio, navigate to the directory that now contains the cloned repo and set the TranslocatoR directory as your working directory. You can now run devtools::install() and TranslocatoR will be installed as normal. Don't forget to run library(TranslocatoR) and then you can then follow the usage example (keeping your working directory set to the cloned TranslocatoR repository folder).

TranslocatoR utilizes the following R packages: "data.table", "gtools", "ggplot2", "stringr", "discreteMTP", "assertthat" and these will be installed automatically if you do not have them on your machine yet.

Below we provide you with a ready-to-use example here, as well as some output-context.

Using TranslocatoR

Execute the following command and make sure to specify non-default arguments if needed.

translocatoR(data.folder, output.folder, samples, options = "pq", binsize = 100000L, cutoff = 0.01, regions = NULL, trfile = NULL, blacklist = T) 

Argument	Comment
data.folder	path to MosaiCatcher data folder
output.folder	absolute path to output folder for TranslocatoR data
samples	samples in MosaiCatcher output folder to be analyzed by TranslocatoR. Use consistent ID throughout the data-set.
options	can take one of multiple values "segments", "pq", "majority"; defaults to "pq".
binsize	which binsize to use, defaults to 100kb
cutoff	cutoff for significant FDR-corrected p-values, defaults to 0.01
regions	list of regions in the format "chr#:-" for potential translocations
trfile	list of manually-identified strand state of suspected translocation
blacklist	List of coordinates for centromeres and short arms for acrocentric chromosomes. Defaults to TRUE and is strongly recommended.

data.folder should be the path to the MosaiCatcher data folder that contains your sample folders of choice. If you decide to use other files please ensure the following path-structure

|__<data.folder>
   |__<sample ID>  
      |__<strandphaser>
         |__<StrandPhaseR_final_output.txt>

options

• pq takes the strand states of the most distal segments on both the p- and q-arms of each chromosome. Translocations are expected to be on the ends of chromosomes.
• majority identifies the majority state (i.e. the most prevalent strand state) of each chromo- some to use for comparison
• segments automatically identifies all recurring segments in a library. This is useful for very complex events.

samples should give distinctive sample-ids. Multiple inputs are possible: samples=c("sampleA", "sampleB")

regions File containing regions of interest in a sample. These regions are the positions where recurrent strand state changes take place, or one expects to see a more complex translocation. TranslocatoR extracts the majority strand state for every given region in each cell and treats the regions as extra segments.

Provide a file in the following format, multiple sample-id inputs are possible:

sample	chrom	start	end
RPE-BM510	chr12	80000	900000
C7	chr5	0	555666

trfile Useful for non-reciprocal translocations: the translo- cated portion of a chromosome will look like a duplication and identifying its state is straightforward.

• Cannot contain more than one sample-id. In order to supply several sample-ids provide several paths for the argument: trfile = c("path1", "path2", ...)
• The first two columns of the .txt file must contain the sample- and cell-ids.
• The states in this file will be treated as an extra segment in each cell.

sample	cell	chr10tr
RPE-BM510	BM510_20306	C
RPE-BM510	BM510_20310	W
RPE-BM510	BM510_20315	W

Output

The following output file structure will be created:

|__<your_output_folder>  
    |__sample-ID                        # creates folder for each given sample
    |   |__haplotypes-per-arm.txt       # raw data output
    |   |__outliers
    |       |__<outliers.txt>           # cells that do not follow the predicted translocation pattern
    |   |__translocations  
    |       |__pvalue-table.txt         # all possible combinations
    |       |__recurrent-segments.txt   # recurring breakpoints that occur > 2x within the data set  
    |       |__translocations.txt       # suggested translocations after applying pvalue cut-off

Some insight on how to understand the given output is given below.

Ready-to-use Example

Here we provide you ready-to-use data for testing TranslocatoR. Run TranslocatoR yourself:

translocatoR(
    data.folder = "data/example-data/example-input", 
    output.folder = "data/example-data/example-output/your-output", 
    samples = "RPE-BM510",
    options = "majority",
    trfile = "data/example-data/example-input/trfile-example.txt",
    )

Input

1. MosaiCatcher output-folder structure containing "RPE-BM510" as a sample: data/example-data/example-input
2. Example file containing the manually-identified strand states of the suspected translocation data/example-data/example-input/trfile-example.txt

Output

1. We provide you with the expected example-run output: data/example-data/example-output/example-run
2. If you have run TranslocatoR yourself the output can be found here: data/example-data/example-output/your-output

Understanding the Output

translocation.txt The primary output-file of interest.

• SegA/SegB: chromosomes that are affected by translocation: SegA translocates to SegB, haplotypespecific.
• cor: Positive/negative correlation.
• p: p-value
• x: Number of matching states.
• n: Number of cells.
• pBH: FDR-adjusted p-value.

The other output files are useful for further reference, but not crucial.

• outliers.txt Outlyer cells that do not match the segregation pattern for any of the identified translocations. Can be used to double-check the listed cells.
• haplotypes-per-arm-txt Matrix containing all split haplotypes per cell.
• pvalue-table.txt Full matrix of segment comparisons and p-values; no cut-off applied.
• recurrent-segments.txt Matrix of recurrent segments (>1x; not entire chr). Lists start-/end-point in N-cells. Can be used to infer breakpoint and thus validate the detected translocations.

Understanding the Example Output

The output-file of primary interest is translocations.txt where TranslocatoR gives a list of detected translocation candidates ordered by FDR-adjusted p-values depending on the set cut-off. This list is affected by the arguments used for the run. In our example you can see of the three candidate translocation partners found for the tested segment (specified by trfile.txt) chr10tr chrX.H2 is the most probable partner. This coincides with the previously described translocation der(X)t(X;10) (Janssen et al., 2011, DOI: 10.1126/science.1210214). Hence, by going back to the data and looking at the candidate translocations in the other files provided by TranslocatoR and the MosaiCatcher pipeline, one can evaluate the candidate translocations.