Release miniprot-0.3 (r137)

lh3 · Sep 22, 2022 · b442b7a · b442b7a
1 parent 457d6de
commit b442b7a
Show file tree

Hide file tree

Showing 4 changed files with 39 additions and 13 deletions.
diff --git a/NEWS.md b/NEWS.md
@@ -1,3 +1,29 @@
+Release 0.3-r137 (22 September 2022)
+------------------------------------
+
+Notable changes:
+
+ * Improvement: fine tune parameters and heuristics: higher non-GT-AG penalty,
+   higher frameshift penalty, higher penalty on in-frame stop codons and a
+   small penalty on long terminal introns. Miniprot is a little more sensitive
+   and a little more accurate, at a minor cost of performance.
+
+ * New feature: richer GFF output. Miniprot now reports per-exon alignment
+   score, number of frameshifts, number of in-frame stop codons and
+   non-canonical donor/acceptor sequences.
+
+ * New feature: added option `--outn` to control the number of alignments per
+   protein to output.
+
+ * New feature: added option `-P` to change the ID prefix in GFF output (#6).
+
+ * Bug fix: fixed a segmentation fault when there are no k-mers on a reference
+   sequence (#4).
+
+(0.3: 22 September 2022, r137)
+
+
+
 Release 0.2-r116 (12 September 2022)
 ------------------------------------
 

diff --git a/README.md b/README.md
@@ -90,20 +90,20 @@ practical for whole genomes. In addition, [Iwata and Gotoh (2012)][spaln2]
 suggest that spaln2 consistently outperforms exonerate, GeneWise, ProSplign and
 genBlastG.
 
-In the evaluation, both miniprot and spaln were set to use 16 CPU threads. We
-used option `-Q7 -O0 -Thomosapi` with spaln. This does global alignment with
-the human-specific splice model.
+In the evaluation, both miniprot-0.3 and spaln-2.4.12 were set to use 16 CPU
+threads. We used option `-Q7 -O0 -Thomosapi` with spaln. This does global
+alignment with the human-specific splice model.
 
 |Metric          |mouse/mp |mouse/sp |chicken/mp|zebrafish/mp|
 |:---------------|--------:|--------:|--------:|--------:|
-|Elapsed time (s)|     347 |   3,714 |     294 |     464 |
+|Elapsed time (s)|     367 |   3,714 |     318 |     494 |
 |# proteins      |  21,844 |  21,844 |  17,007 |  30,313 |
-|# mapped        |  19,253 |  18,847 |  13,284 |  19,797 |
-|# single-exon   |   2,878 |         |   1,110 |   1,857 |
-|# predicted junc| 164,718 | 173,475 | 131,346 | 176,044 |
-|# non-ovlp junc |     402 |   1,490 |     482 |     960 |
-|# confirmed junc| 157,400 | 162,303 | 117,416 | 151,912 |
-|% confirmed     |    95.6 |    93.6 |    89.4 |    86.3 |
+|# mapped        |  19,375 |  18,847 |  13,496 |  20,382 |
+|# single-exon   |   2,960 |         |   1,253 |   2,098 |
+|# predicted junc| 165,219 | 173,475 | 132,275 | 179,961 |
+|# non-ovlp junc |     377 |   1,490 |     435 |     800 |
+|# confirmed junc| 158,164 | 162,303 | 119,492 | 157,004 |
+|% confirmed     |    95.7 |    93.6 |    90.3 |    87.2 |
 
 On the human-mouse dataset, miniprot finds fewer novel splice junctions,
 implying higher specificity, but spaln finds more confirmed junctions, implying

diff --git a/miniprot.1 b/miniprot.1
@@ -1,4 +1,4 @@
-.TH miniprot 1 "21 September 2022" "miniprot-0.2-dirty (r133)" "Bioinformatics tools"
+.TH miniprot 1 "22 September 2022" "miniprot-0.3 (r137)" "Bioinformatics tools"
 .SH NAME
 .PP
 miniprot - protein-to-genome alignment with splicing and frameshifts
@@ -153,7 +153,7 @@ Parent	str	identifier of the parent feature
 Identity	real	fraction of exact amino acid matches
 Positive	real	fraction of positive amino acid matches
 Donor	str	2bp at the donor site if not GT
-Accepter	str	2bp at the acceptor site if not AG
+Acceptor	str	2bp at the acceptor site if not AG
 Frameshift	int	number of frameshift events in alignment
 StopCodon	int	number of in-frame stop codons
 Target	str	Protein coordinate in alignment

diff --git a/miniprot.h b/miniprot.h
@@ -3,7 +3,7 @@
 
 #include <stdint.h>
 
-#define MP_VERSION "0.2-r136-dirty"
+#define MP_VERSION "0.3-r137"
 
 #define MP_F_NO_SPLICE    0x1
 #define MP_F_NO_ALIGN     0x2