Hello! I am encountering a similar problem using HTSeq-0.5.4p2 I tried to count reads of Arabidopsis (TAIR10_GFF3_genes.representative.gmap.gff ) with the " -m union -s no " options using a .sam file produced by GSNAP. The data is paired end 101bp unstranded reads by illumina HiSeq. The .sam files have been used previously by htseq-count to map on the same gff file using the "-m intersection-strict" , and everything worked fine. However I do not know which htseq-count version was used as it was done within a Galaxy environment where I do not know the version, but most likely it comes from 2012. Also when I use only a short test .sam file everything works fine. The program terminates with an error after processing a substantial number of line pairs (for different .sam files it is a different number, probably depends on the occurrence of a problematic read(?)): ... 4100000 sam line pairs processed. Error occured in line 8217724 of file total1.sam. Error: Cannot subset to zero-length interval. [Exception type: IndexError, raised in _HTSeq.pyx:348] I tried the same with HTSeq-0.5.4.p1 but the error was identical. I use python2.7, ubuntu 11.10. I would be grateful for help Thanks matyas [[alternative HTML version deleted]]