Skip to content
samtools release 1.9:

 * Samtools mpileup VCF and BCF output is now deprecated.  It is
   still functional, but will warn.  Please use bcftools mpileup
   instead. (#884)

 * Samtools mpileup now handles the '-d' max_depth option
   differently.  There is no longer an enforced minimum, and '-d
   0' is interpreted as limitless (no maximum - warning this may
   be slow).  The default per-file depth is now 8000, which
   matches the value mpileup used to use when processing a single
   sample.  To get the previous default behaviour use the higher
   of 8000 divided by the number of samples across all input
   files, or 250. (#859)

 * Samtools stats new features:

   - The '--remove-overlaps' option discounts overlapping portions
     of templates when computing coverage and mapped base counting.
     (#855)



   - When a target file is in use, the number of bases inside the
     target is printed and the percentage of target bases with
     coverage above a given threshold specified by the
     '--cov-threshold' option. (#855)

   - Split base composition and length statistics by first and last
     reads. (#814, #816)

 * Samtools faidx new features:

   - Now takes long options. (#509, thanks to Pierre Lindenbaum)

   - Now warns about zero-length and truncated sequences due to the
     requested range being beyond the end of the sequence. (#834)



   - Gets a new option (--continue) that allows it to carry on when a
     requested sequence was not in the index. (#834)

   - It is now possible to supply the list of regions to output in a
     text file using the new '--region-file' option. (#840)

   - New '-i' option to make faidx return the reverse complement of
     the regions requested. (#878)

   - faidx now works on FASTQ (returning FASTA) and added a new fqidx
     command to index and return FASTQ. (#852)

 * Samtools collate now has a fast option '-f' that only operates on
   primary pairs, dropping secondary and supplementary.  It tries to
   write pairs to the final output file as soon as both reads have
   been found. (#818)

 * Samtools bedcov gets a new '-j' option to make it ignore deletions
   (D) and reference skips (N) when computing coverage. (#843)

 * Small speed up to samtools coordinate sort, by converting it to
   use radix sort. (#835, thanks to Zhuravleva Aleksandra)

 * Samtools idxstats now works on SAM and CRAM files, however this
   isn't fast due to some information lacking from indices. (#832)

 * Compression levels may now be specified with the level=N
   output-fmt-option.  E.g. with -O bam,level=3.

 * Various documentation improvements.

 * Bug-fixes:

   - Improved error reporting in several places. (#827, #834, #877,
     cd7197)

   - Various test improvements.

   - Fixed failures in the multi-region iterator (view -M) when
     regions provided via BED files include overlaps (#819,
     reported by Dave Larson).

   - Samtools stats now counts '=' and 'X' CIGAR operators when
     counting mapped bases. (#855)

   - Samtools stats has fixes for insert size filtering (-m, -i).
     (#845; #697 reported by Soumitra Pal)

   - Samtools stats -F now longer negates an earlier -d option.
     (#830)

   - Fix samtools stats crash when using a target region. (#875,
     reported by John Marshall)

   - Samtools sort now keeps to a single thread when the -@ option is
     absent. Previously it would spawn a writer thread, which could
     cause the CPU usage to go slightly over 100%. (#833, reported by
     Matthias Bernt)

   - Fixed samtools phase '-A' option which was incorrectly defined
     to take a parameter. (#850; #846 reported by Dianne Velasco)

   - Fixed compilation problems when using C_INCLUDE_PATH. (#870;
     #817 reported by Robert Boissy)

   - Fixed --version when built from a Git repository. (#844, thanks
     to John Marshall)

   - Use noenhanced mode for title in plot-bamstats.  Prevents
     unwanted interpretation of characters like underscore in
     gnuplot version 5. (#829, thanks to M. Zapukhlyak)

   - blast2sam.pl now reports perfect match hits (no indels or
     mismatches). (#873, thanks to Nils Homer)

   - Fixed bug in fasta and fastq subcommands where stdout would not
     be flushed correctly if the -0 option was used.

   - Fixed invalid memory access in mpileup and depth on alignment
     records where the sequence is absent.