Hello!
I am trying to run the program SV-bay (https://github.com/InstitutCurie/SV-Bay, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4896370/) in order to find translocations in our 60x tumour + 30x normal WGS datasets. However, Im having issues getting the second step main_probabilities.py) to run. In the log file, it is complaining about the gem files and the fasta files being of different lengths:
main_probabilities.py line:197 INFO [2016-11-11 11:12:22,971] Chromosomes to process:['chr1', 'chr2', 'chr3', 'chr4', 'chr5', 'chr6', 'chr7', 'chr8', 'chr9', 'chr10', 'chr11', 'chr12', 'chr13', 'chr14', 'chr15', 'chr16', 'chr17', 'chr18', 'chr19', 'chr20', 'chr21', 'chr22', 'chrX', 'chrY']
bayesianinput.py line:48 INFO [2016-11-11 11:12:22,971] Loading data...
bayesianinput.py line:106 INFO [2016-11-11 11:12:22,971] Loading normal fragments...
bayesianinput.py line:137 INFO [2016-11-11 11:32:03,126] Loading links...
bayesianinput.py line:158 INFO [2016-11-11 11:32:21,881] Constructing sublinks...
bayesianinput.py line:223 WARNING [2016-11-11 11:32:32,877] Lambda file not found, calculating lambda
bayesianinput.py line:252 INFO [2016-11-11 11:32:32,877] Started calculating lambda_gen
bayesianinput.py line:67 WARNING [2016-11-11 11:32:39,167] Lengths of .fa and .gem sequences do not match for chrom chr1 (249250626 and 248956422)
bayesianinput.py line:69 WARNING [2016-11-11 11:32:39,167] Some flanking regions will be truncated to match the minimal length and some links (intersecting with chrom tail) will be dropped.
bayesianinput.py line:70 WARNING [2016-11-11 11:32:39,167] Please fix your .gem file.
Which is interesting because I am using the hg19 fasta and gem files provided on the GitHub page (see link above). The actual python error that the program is printing out looks as follows:
[x@bespin sv-bay-data]$ /apps/CLC_ExternalApps/SV-Bay/miniconda2/bin/python2.7 -B /apps/CLC_ExternalApps/SV-Bay/SV-Bay/src/main_probabilities.py -c /apps/CLC_ExternalApps/SV-Bay/SV-Bay/config/config.yaml
Traceback (most recent call last):
File "/apps/CLC_ExternalApps/SV-Bay/SV-Bay/src/main_probabilities.py", line 199, in <module>
input_data = BayesianInputData(config, stats, chromosomes)
File "/apps/CLC_ExternalApps/SV-Bay/SV-Bay/src/bayesianinput.py", line 57, in __init__
self.__LoadLambda(config, stats)
File "/apps/CLC_ExternalApps/SV-Bay/SV-Bay/src/bayesianinput.py", line 224, in __LoadLambda
(lambda_norm,lambda_abnorm) = self.__CalculateLambda(config, stats)
File "/apps/CLC_ExternalApps/SV-Bay/SV-Bay/src/bayesianinput.py", line 363, in __CalculateLambda
if gem_val_abnormal[i/step]!=-1:
IndexError: list index out of range
I have successfully run through all the steps when doing the analysis on only a single chromosome (chr17). If anyone has a clue about what the problem is or how I can recreate the gem mappability files, I would be deeply grateful.
/A
I had issues with the clustering step too. The thing that solved it was to use the exact versions of of the prerequisites there were out when the last commit of SV-bay was done: numpy==1.8.2 scipy==0.13.3 matplotlib==1.3.1 pyaml==16.9.0 pysam==0.9.1.4 joblib==0.10.2
I installed a separate python via miniconda and put these exact versions into it.
can you give me your test data ? I wonder to know the reason and different of my data.
Sorry, I cannot give you my test data because it is from patients.