Bug for reading the tissue_positions.csv

[tungtp99]

Report

squidpy/src/squidpy/read/_read.py

Line 83 in 538427b

header=1 if tissue_positions_file.name == "tissue_positions.csv" else None,

Should we use header=0 instead? header=1 make the data loose information of the first spot after change the column name

Thank you!

Versions

[selmanozleyen]

Hi thanks for pointing this out! Which dataset did you observe this in? Does this happen also in 10x Genomics Visium formatted datasets?

[shashkat]

I checked in this 10x dataset: https://www.10xgenomics.com/datasets/visium-v2-human-lymph-node, and it looks that the author's point is correct. The tissue_positions.csv file has a header row.

$ head tissue_positions.csv
barcode,in_tissue,array_row,array_col,pxl_row_in_fullres,pxl_col_in_fullres
GTCACTTCCTTCTAGA-1,1,0,0,23891.771232682222,39028.97429208872
CACGGTCTCCTTACGA-1,1,0,2,23897.618791707442,38663.634433938605
ATAGCTGCGGATAAGA-1,1,0,4,23903.46462886722,38298.2952273851
GTCAGTATGTCCGGCG-1,1,0,6,23909.31218148734,37932.95571104603
ATGTACCAGTTACTCG-1,1,0,8,23915.159740512565,37567.615852895906
ACGCTCAGTGCACCGT-1,1,0,10,23921.00558727999,37202.276133625826
TCACTAACGTATAGTT-1,1,0,12,23926.853130292468,36836.93713000333
CGGTTAGGCCTGGACG-1,1,0,14,23932.699492979013,36471.59724943537
GATATCACCAGCATGG-1,1,0,16,23938.546539287658,36106.257381677606

Hence, it would be correct to use header=0 instead of header=1.

I want to submit a PR for the same. If you don't mind, I will go ahead and submit one within next few days.