The following folders include both the source data and the scripts used to plot them. 

DGs_5methods: CRSSANT-assembled DGs for the 5 crosslink-ligation methods analyzed in this paper. 
Fig. 1. N/A
Fig. 2A-B. bed file for human rRNA and snRNA helices
Fig. 2GHIJ. alignment counts for the benchmarking of 4 methods on simulated reads. 
Fig. 2K. alignment counts for the benchmarking of 4 methods on real crosslink-ligation data
Fig. 2LM. BAM file for the alignments. 
Fig. 3E. summary of alignment numbers for the crosslink-ligation data using optimized STAR
Fig. 4EF. performance of CRSSANT on simulated DGs. 
Fig. 4G. complete datasets are in Supplemental Fig. 5C-D
Fig. 4HIJ. human U2 snRNA model and PARIS data from human and mouse cells
Fig. 4K. human U2 SHARC data
Fig. 5E. 28S rRNA TG coverage before and after shuffling. 
Fig. 5F. 28S rRNA DG and TG coverages before and after shuffling
Fig. 6CDE. RNA homodimers data. 

Supplemental Fig. 1. benchmarking 4 mapping methods on simulated reads. 
Supplemental Fig. 2EF. nucleotide bias in short gaps
Supplemental Fig. 2G. benchmarking 4 mapping methods on real crosslink-ligation reads.  
Supplemental Fig. 3. N/A
Supplemental Fig. 4. N/A
Supplemental Fig. 5C. benchmarking CRSSANT on simulated alignments, cliques algorithm
Supplemental Fig. 5D. benchmarking CRSSANT on simulated alignments, spectral algorithm
Supplemental Fig. 6A-D. comparing simulated and CRSSANT-assembled DGs. 
Supplemental Fig. 7A. simulated DG BAM files with varying coverages
Supplemental Fig. 7B. numbers on the right of the plot
Supplemental Fig. 7C. numbers on the right of the plot
Supplemental Fig. 7D. wall clock time for CRSSANT clustering 
Supplemental Fig. 8G. human U3 SHARC data
Supplemental Fig. 9. testing CRSSANT on the rRNA using published crosslink-ligation data
Supplemental Fig. 10. gapm more data other RNAs. 
Supplemental Fig. 11. N/A
Supplemental Fig. 12. data are in the fig. 6 folder
Supplemental Fig. 13. picornavirus homodimers
Supplemental Fig. 14C. quantification of the homodimer RT-PCR. 
