CNVkit - ImportError: cannot import name 'Int64Index' from 'pandas'

[pvvhoang]

Dear Dr. Eric,

I would like to use CNVkit for identifying copy number amplifications from whole genome sequencing data. I have installed CNVkit on our server and run it; however, I have experienced the below error in the step of running the example workflow.

//==================================
(ecDNA) [vp8928@gadi-login-09 test]$ make
cnvkit.py batch -n -f formats/chrM-Y-trunc.hg19.fa -m wgs formats/na12878-chrM-Y-trunc.bam -d build
Traceback (most recent call last):
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/bin/cnvkit.py", line 4, in
from cnvlib import commands
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/cnvlib/init.py", line 1, in
from skgenome.tabio import write
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/skgenome/init.py", line 1, in
from . import tabio
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/skgenome/tabio/init.py", line 13, in
from ..gary import GenomicArray as GA
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/skgenome/gary.py", line 9, in
from .intersect import by_ranges, into_ranges, iter_ranges, iter_slices
File "/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/skgenome/intersect.py", line 11, in
from pandas import Int64Index
ImportError: cannot import name 'Int64Index' from 'pandas' (/g/data/ur22/vp8928/miniconda3/envs/ecDNA/lib/python3.11/site-packages/pandas/init.py)
make: *** [Makefile:49: build/na12878-chrM-Y-trunc.bintest.cns] Error 1
//===================================

May you have a review and give me your advice? Thank you for your time and consideration of my request.

Best regards,

[SevenTea7]

I also encountered the same problem. When I opened /pandas/init.py, I found that the.py file was 0kb, and I also reported an error when I re-installed this package.
(ampsuite) liuhui@B1:~/ysr/app/cnvkit-0.9.7$ /data/liuhui/ysr/app/cnvkit-0.9.7/cnvkit.py Traceback (most recent call last): File "/data/liuhui/ysr/app/cnvkit-0.9.7/cnvkit.py", line 4, in <module> from cnvlib import commands File "/data/liuhui/ysr/app/cnvkit-0.9.7/cnvlib/__init__.py", line 1, in <module> from skgenome.tabio import write File "/data/liuhui/ysr/app/cnvkit-0.9.7/skgenome/__init__.py", line 1, in <module> from . import tabio File "/data/liuhui/ysr/app/cnvkit-0.9.7/skgenome/tabio/__init__.py", line 13, in <module> from ..gary import GenomicArray as GA File "/data/liuhui/ysr/app/cnvkit-0.9.7/skgenome/gary.py", line 9, in <module> from .intersect import by_ranges, into_ranges, iter_ranges, iter_slices File "/data/liuhui/ysr/app/cnvkit-0.9.7/skgenome/intersect.py", line 11, in <module> from pandas import Int64Index ImportError: cannot import name 'Int64Index' from 'pandas' (/data/liuhui/anaconda3/envs/ampsuite/lib/python3.10/site-packages/pandas/__init__.py)

[jluebeck]

+1 to this issue.

It appears the Int64Index import was deprecated in pandas 1.4.0, and removed in 2.0.0.
https://pandas.pydata.org/docs/whatsnew/v1.4.0.html#deprecated-int64index-uint64index-float64index

https://pandas.pydata.org/docs/whatsnew/v2.0.0.html#index-can-now-hold-numpy-numeric-dtypes

Edit: I believe this is resolved in cnvkit 0.9.10 and later