TypeError: 'numpy.float64' object cannot be interpreted as an integer

[David Castillo Rodríguez]

Hello everyone,

I'm running DANPOS3. I actually created a Docker container to run this program using the following commands:

FROM python:latest

WORKDIR /batchx

RUN apt-get update && apt install -y r-base && apt -y autoremove

RUN git clone https://github.com/sklasfeld/DANPOS3.git

RUN cd DANPOS3 && pip install -r requirements.txt

RUN pip install scipy

I ran dpos with a single BAM file sample and everything went well. But every time I run dpos with more than one sample I get the same error:

fine-tuning positions for tmp_A_972-t0-20U-n3_S17_L003.mini_A1.sorted by comparing to the reference map...

Traceback (most recent call last):

  File "/batchx/DANPOS3/danpos.py", line 1365, in <module>

    if sys.argv[1]=='dpos':runDANPOS(command='dpos')

  File "/batchx/DANPOS3/danpos.py", line 396, in runDANPOS

    danpos(\

  File "/batchx/DANPOS3/functions.py", line 454, in danpos

 pooledgroups[groupname].fillPositions(dic=tdic,file=os.path.join(opath,'pooled',groupname+addname+"positions.ref_adjust.xls"),width=width,distance=distance,edge=edge,pcut=pheight,height=height,calculate_P_value=1,mode='w',title_line=1,poscal=1,rd=rd)

  File "/batchx/DANPOS3/wig.py", line 492, in fillPositions

    smts.fetchValueFromWig(twig)

  File "/batchx/DANPOS3/summits.py", line 43, in fetchValueFromWig

    if tp>wig.data[cr].size:wig.data[cr].resize(tp+1,refcheck=0)

TypeError: 'numpy.float64' object cannot be interpreted as an integer

In this last attempt, this is the command I was using:
python3 /batchx/DANPOS3/danpos.py dpos /tmp/A/972-t0-20U-n3_S17_L003.mini_A1.sorted.bam,/tmp/A/972-t0-20U-n3_S17_L003.mini_A2.sorted.bam:/tmp/condition-b/972-t0-20U-n3_S17_L003.mini_B1.sorted.bam,/tmp/condition-b/972-t0-20U-n3_S17_L003.mini_B2.sorted.bam --paired 1 --height 5 --out /batchx/output/dpos/

I was trying to compare two conditions, with two replicates each. 

All BAM files were created using bowtie2. All BAM files have paired reads and were sorted and indexed, before running dpos. Hope somebody knows what raises this error. The most I'm able to infer is that a float is being used where an int was expected.

If more info is required, let me know!

Cheers!

David

[David Castillo Rodríguez]

Adding the output from running the print_versions.py script:

# python3 /batchx/DANPOS3/print_versions.py

Python version

3.9.0 (default, Oct 13 2020, 20:14:06) 

[GCC 8.3.0]

-------------------------------------------------

R version 3.5.2 (2018-12-20) -- "Eggshell Igloo"

Copyright (C) 2018 The R Foundation for Statistical Computing

Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.

You are welcome to redistribute it under the terms of the

GNU General Public License versions 2 or 3.

For more information about these matters see

http://www.gnu.org/licenses/.

0

-------------------------------------------------

samtools 1.9

Using htslib 1.9

Copyright (C) 2018 Genome Research Ltd.

0

-------------------------------------------------

AVAILABLE python packages:

numpy: 1.18.5

rpy2: 3.3.3

pysam: 0.16.0.1

argparse: 1.1

Great job! No missing python packages!

[David Castillo Rodríguez]

I managed to avoid the summits.py script from crashing by modifying line 43.

from this:

if tp>wig.data[cr].size:wig.data[cr].resize(tp+1,refcheck=0)

to this:

if tp>wig.data[cr].size:wig.data[cr].resize(int(tp)+1,refcheck=0)

I just coerced tp to an integer. This avoids the script from crashing but not sure if this is a valid solution.

Please let me know if this solution is appropriate.

Best wishes,

David

[sjk...@gmail.com]

Thank you so much for your feedback. When I converted the scripts from python2 to python3 this bug was most likely born. The two types treat numerical objects differently. This is the solution. Thanks for your help!

Sincerely,

Sammy

[David Castillo Rodríguez]

Hello Sammy,

happy to help. Unfortunately, soon after, I stumbled into a new problem: https://groups.google.com/g/danpos/c/J4T_va8XrE4
I decided to take a step back and use version 2.2.2. So far it has been working flawlessly : )

Cheers!

David