OptiType crashing with message "this is indicative of a SERIOUS ERROR in the expression reuse detection scheme."

[alexanderwhatley]

OptiType is crashing with the following error message from the log:

mapping with 8 threads...

 0:00:59.80 Mapping PGDX3144N_HLA_filtered_fished.fastq to GEN reference...

 0:03:48.28 Generating binary hit matrix.
0:03:48.31 Loading ../OptiType_output/2017_03_07_16_43_12/2017_03_07_16_43_12_1.bam started. Number of HLA reads loaded (updated every thousand):
1K...2K...3K...4K...5K...6K...7K...8K...9K...10K...11K...12K...13K...14K...
 0:04:18.66 14383 reads loaded. Creating dataframe...
0:04:21.58 Dataframes created. Shape: 14383 x 11179, hits: 4063907 (4063907), sparsity: 1 in 39.56

 0:04:37.63 temporary pruning of identical rows and columns

 0:04:39.04 Size of mtx with unique rows and columns: (767, 1332)
0:04:39.04 determining minimal set of non-overshadowed alleles

 0:04:55.87 Keeping only the minimal number of required alleles (168,)

 0:04:55.91 Creating compact model...

starting ilp solver with 1 threads...

 0:04:58.69 Initializing OptiType model...
ERROR: Rule failed when generating expression for objective read_cov:
    RuntimeError: Expression entered generate_expression() with too few references (-1<0); this is indicative of a SERIOUS ERROR in the expression reuse detection scheme.
ERROR: Constructing component 'read_cov' from data=None failed:
    RuntimeError: Expression entered generate_expression() with too few references (-1<0); this is indicative of a SERIOUS ERROR in the expression reuse detection scheme.

The explicit error message is:

Traceback (most recent call last):
  File "OptiTypePipeline.py", line 404, in <module>
    config.get("ilp", "solver"), threads, verbosity=VERBOSE)
  File "/n/regal/nowak_lab/immunotherapy/OptiType/model.py", line 97, in __init__
    model.reconst[a] * model.x[a] for a in model.L), sense=maximize)
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/block.py", line 484, in __setattr__
    self.add_component(name, val)
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/block.py", line 890, in add_component
    val.construct(data)
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/objective.py", line 307, in construct
    tmp = _init_rule(_self_parent)
  File "/n/regal/nowak_lab/immunotherapy/OptiType/model.py", line 96, in <lambda>
    rule=lambda model: sum(model.occ[r] * (model.y[r] - model.beta * (model.re[r])) for r in model.R) - sum(
  File "/n/regal/nowak_lab/immunotherapy/OptiType/model.py", line 96, in <genexpr>
    rule=lambda model: sum(model.occ[r] * (model.y[r] - model.beta * (model.re[r])) for r in model.R) - sum(
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/numvalue.py", line 460, in __sub__
    return generate_expression(_sub,self,other)
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/expr_coopr3.py", line 1028, in generate_expression
    other = _generate_expression__clone_if_needed(other, 0)
  File "/n/home05/aewhatley/anaconda3/lib/python3.6/site-packages/pyomo/core/base/expr_coopr3.py", line 918, in _generate_expression__clone_if_needed
    % ( getrefcount(obj) - UNREFERENCED_EXPR_COUNT, ))
RuntimeError: Expression entered generate_expression() with too few references (-1<0); this is indicative of a SERIOUS ERROR in the expression reuse detection scheme.

Do you know what the problem could be?

[andras86]

Not really, haven't seen this error before. Do other samples run through fine? If so, perhaps you could e-mail me the intermediate bam file so I can try to reproduce the error. It's deleted by default so make sure you set deletebam=false in your config file. Also, what solver do you use?

[alexanderwhatley]

Just emailed you.

[messersc]

Hi guys,

apparently pyomo does not work with Python 3.6 (yet): http://stackoverflow.com/a/42117803 So if you downgrade to python 3.5 this error should go away.

HTH, Clemens

[alexanderwhatley]

Closed.