This add-on of TADbit allows you to run Loopbit, a Convolutional Neural Network (CNN) able to identify chromatin loops from 3C-based experiments [CITE PAPER].
Loopbit was written and tested using Python 3.6, 3.7.
It requires the packages matplotlib (3.2.1), scipy (1.0.0), numpy (1.14.5), tensorflow (1.10.0), and seaborn (0.9.0) and tqdm (4.32.2).
In order to install the package you can download the code from our GitHub repo and install it manually. If needed, the dependencies will be downloaded automatically; if you encounter problems, try to install the other requirements manually.
git clone https://github.com/3DGenomes/loopbit # or download manually
cd loopbit
python setup.py installAfter the installation, you can run the provided example to familiarize with the functions of loopbit.
Use loopbit -h for quick help and orientation.
For the tutorial we are going to use a region of the chromosome 21 from the replicate H of . We are going to analyze a 1 Mb region, from 33,379,694 to 34,327,701.
loopbit scan \
example/chr21_GM12878_replicateH.tsv \
chr21 \
45000-3000000 \
5000 \
9 \
3 \
example/ \It will generate a output file with 2 columns, with a label and its probability of being a loop. The label will be chromosome_binI_binJ:
chr21_562183_562202 1.7374185596594756e-19
chr21_562186_562205 1.4961456010730012e-13
In order to visualize the results we will run the following command-line:
loopbit plot \
example/chr21_GM12878_replicateH.tsv \
example/loopbit_results_chr21.tsv \
33379694-34327701 \
5000 \
example/ \This is the plot that will be displayed, containing the contact information in the upper triangle, and the probabilities cloud in the lower triangle:
Loopbit has these two main commands:
scanto scan the region of interest or the full chromosome to obtain the loop prediction.plotto generate a plot of a region of interest with the contact matrix and the cloud of probabilities.
First, the region of interest or the full chromosome have to be scanned, thus obtaining the loop prediction value. To run scan, the following arguments are needed:
- A normalized Hi-C matrix file in a TSV format. This should be tab-delimited, where each line has three columns:
row index column index value
- The name of the chromosome or a specific label to use to store results.
- Interval of distance to check the presence of chromatin loops (bp). For example, if we want to inspect from 45kb to 3Mb, we will put 45000-3000000.
- Resolution of the experiment to inspect, in bp.
- Number of bins considered to check if contains a loop. For example, if we put 9, submatrices of 9x9 will be used to check if contains a loop.
NOTE: Recommended to have around 9 bins, no less than 3
- Step size to scan, if 1, it will scan all the chromosome moving +1bin each time
- Directory to store the results.
This function allows to observe the results from a region of interest.
- A normalized Hi-C matrix file in a TSV format. This should be tab-delimited, where each line has three columns:
row index column index value
- The outplut file from loopbit.
- Region of interest to be plotted with the following format, in bp: start-end.
- Resolution of the experiment to inspect, in bp.
- Directory to store the results.
NOTE: there is the parameter -G or gaussian that can be changed if needed. By default is 4. Here you can learn more about how this filter works: https://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.gaussian_filter.html
Check the label FAQ in TADbit issues.
If your question is still unanswered feel free to open a new issue.
This add-on of TADbit is currently developed at the MarciusLab with the contributions of Silvia Galan and François Serra.
Please, cite this article if you use TADbit.
Serra, F., Baù, D., Goodstadt, M., Castillo, D. Filion, G., & Marti-Renom, M.A. (2017).
Automatic analysis and 3D-modelling of Hi-C data using TADbit reveals structural features of the fly chromatin colors.
PLOS Comp Bio 13(7) e1005665. doi:10.1371/journal.pcbi.1005665 <https://doi.org/10.1371/journal.pcbi.1005665>_