Why does cellranger aggr throw an error "Your header row is missing a required field: library_id."?

Question: Why does the cellranger aggr pipeline complain that I am missing the library_id field in the header row when I view it in my text editor?

Answer: There are at least two possible causes of the error:

(1) One possible explanation is the Unicode byte order mark (BOM), which is represented by a hidden character <U+FEFF>. Certain programs such as Excel and the Linux cat command will not display the character, so a CSV file appears to conform to formatting requirements. However, cellranger aggr will crash while parsing such a CSV file.

It is possible to check the header line for BOM on Linux using the following command, and quit by pressing 'q':

less Aggregation.csv

It is also possible to fix the CSV file by running the following series of Linux commands:

# Erase <U+FEFF> from file, save result to tmp
awk '{ gsub(/\xef\xbb\xbf/,""); print }' Aggregation.csv > tmp
# Rename tmp to Aggregation.csv
mv tmp Aggregation.csv