READS ARE NOT AT LEAST 75 BP LONG error

[cooketho]

ORP fails on the test data set when I trim the reads to 75 bp, with the following error:

IT LOOKS LIKE YOUR READS ARE NOT AT LEAST 75 BP LONG,
 PLEASE EDIT YOUR COMMAND USING THE SPADES2_KMER=INT FLAGS,
 SETTING THE ASSEMBLY KMER LENGTH LESS THAN YOUR READ LENGTH 

/bin/bash: line 8: shell: command not found
make: *** [readcheck] Error 127

The culprit seems to be the following line from oyster.mk:

if [ $$(gzip -cd $${READ1} | head -n 400 | awk '{if(NR%4==2) {count++; bases += length} } END{print int(bases/count)}') -gt 75 ] && [ $$(gzip -cd $${READ2} | head -n 400 | awk '{if(NR%4==2) {count++; bases += length} } END{print int(bases/count)}') -gt 75 ];\

After some digging, it looks like rnaspades.py actually does need k-mer -k to be an odd number less than the read length. This requirement for an odd number should also be added to the error message, which should read "76 bp", not "75 bp". Furthermore, the spades developers evidently recommend -k of about half the read length (ablab/spades#215). There should probably be some discussion of this in the ORP documentation.

My workaround is to change that line to:

if [ $$(gzip -cd $${READ1} | head -n 400 | awk '{if(NR%4==2) {count++; bases += length} } END{print int(bases/count)}') -gt ${SPADES2_KMER} ] && [ $$(gzip -cd $${READ2} | head -n 400 | awk '{if(NR%4==2) {count++; bases += length} } END{print int(bases/count)}') -gt ${SPADES2_KMER} ];\

so that "75" is not hard-coded in there.

[macmanes]

Hi,

Thanks for your reports - I will respond to each. In this case, the error message has been updated to 76bp, and I will add the part about odd numbers. I know SPAdes rec. is 50% of read length, but we with this second SPAdes assembly we are trying to to recover specifically higher expression transcripts.

Of note, you could have avoided editing the code by passing the SPADES2_KMER=INT flag, if I'm understanding your issue properly.

[cooketho]

I could be wrong, but I don't think the issue can be fixed by just passing the SPADES2_KMER=INT flag. I tried this, but couldn't get it to work. The reason is: if that if statement executes (which it does if your reads are 75 bp, regardless of what you set SPADES2_KMER as), it prints the error message and then calls $$(shell exit);\, ending the program. The if statement should check whether the read length is compatible with the current SPADES2_KMER setting, not whether it is > 75.