Heng Li
b58b97423a
r1281: with --write-junc, ignore aln with low mapQ
2025-04-17 21:46:54 -04:00
Heng Li
df9e650346
r1280: heuristic to avoid aligning full introns
...
This speeds up alignment a lot
2025-04-16 23:38:39 -04:00
Heng Li
7a540c37ca
r1278: reduced --min-dp-len to 20 for splice:sr
...
30% reduced time at 0.01% more junction errors
2025-04-16 21:55:59 -04:00
Heng Li
c19e3ccb86
r1277: don't apply rescored filtering with -P
...
Since v2.19-ish, minimap2 rescores base alignment based on the best alignment
of a read. This heuristic sometimes improve the mapping accuracy of the best
alignment but may too aggressively filter weaker hits.
Resolve #969
2025-04-14 22:26:47 -04:00
Heng Li
94d171b01e
r1276: skip unnecessary reverse spliced alignment
...
for short-read RNA-seq only
2025-04-14 14:57:23 -04:00
Heng Li
01ccedd5a0
r1275: renamed --jump-pass1 to --pass1
...
Also documented 2-pass
2025-04-13 17:07:39 -04:00
Heng Li
819b3bf017
r1274: a little better pass-1
...
Also fixed a bug in jump
2025-04-13 13:55:15 -04:00
Heng Li
e88110463a
r1273: reading pass1 junctions works
...
but we need extra logic when using them. Tomorrow.
2025-04-13 00:47:28 -04:00
Heng Li
fb81e150f2
r1272: support pass-1 junction processing
...
NOT tested yet
2025-04-12 23:15:04 -04:00
Heng Li
d930ea94ad
r1271: code refactoring in prep for 2-pass
2025-04-11 17:20:34 -04:00
Heng Li
a832a42f6f
r1270: fixed ts tag for PE reads
...
this also fixed the wrong --write-junc
2025-04-11 01:08:16 -04:00
Heng Li
a5411fc3c0
r1269: added --write-junc
...
in preparation for 2-pass
2025-04-11 01:00:15 -04:00
Heng Li
bd03d975fc
r1268: avoid extra small introns
2025-04-07 23:32:56 -04:00
Heng Li
9c3c4b1ce8
r1267: penalize introns without signals
2025-04-07 09:51:33 -04:00
Heng Li
3542a3d153
r1266: fine tune mapq for short RNA-seq reads
2025-04-07 00:57:39 -04:00
Heng Li
75619c7b51
r1265: prefer spliced alignment
...
mapq needs to be elevated
2025-04-07 00:31:04 -04:00
Heng Li
fbb9c0fcba
r1264: added --jump-min-match, default to 3
...
To match STAR
2025-04-06 23:18:55 -04:00
Heng Li
af094640e5
r1263: ~5-10% performance improvement
...
Via larger batches and more short-read heuristics. Identical alignment on 2
million reads. Short DNA-seq read alignment may be improved in corner cases.
2025-04-06 20:48:55 -04:00
Heng Li
38acd6617f
r1261: code clean up; renamed --jump-bed to -j
...
Also added --pairing to replace --no-pairing and --pe-ind-chain
2025-04-06 18:32:58 -04:00
Heng Li
54a4718c9b
r1259: --jump-bed now functional
2025-04-06 17:12:49 -04:00
Heng Li
dbc12b2838
r1258: update blen, mlen and dp_max0
2025-04-06 15:26:01 -04:00
Heng Li
2ed264db4e
r1257: fixed the sorting of BED files
2025-04-06 10:53:16 -04:00
Heng Li
a8094ad859
r1256: working for one example, but still buggy
2025-04-06 10:33:57 -04:00
Heng Li
1877818239
r1255: moved jump code to a separate file
2025-04-06 09:33:51 -04:00
Heng Li
9ede5c4255
backup
2025-04-06 00:07:15 -04:00
Heng Li
a8c567b5e9
r1251: read junctions for jumps
2025-04-04 16:23:19 -04:00
Heng Li
cbe8d61ca4
r1249: remove redundant junctions in --junc-bed
2025-04-03 20:43:06 -04:00
Heng Li
9d06cef13e
r1248: support --end-bonus in splice mode
...
but this is not enabled by default for now
2025-04-02 21:39:41 -04:00
Heng Li
bbf0bb871b
r1245: allow shorter hits for splice:sr
2025-04-02 00:13:37 -04:00
Heng Li
83e9b2e28c
r1242: append /[12] to read name in --frag mode
...
Resolve #1079
2025-04-01 10:48:15 -04:00
Heng Li
e816fd071c
r1241: error out on unintended --score-N
...
Resolve #1226
2025-04-01 10:12:53 -04:00
Heng Li
54fa925e2e
r1240: fixed wrong logging information
...
resolve #1192
2025-04-01 09:17:39 -04:00
Heng Li
d4a396c5c3
r1239: resovled #963
2025-03-31 23:15:34 -04:00
Heng Li
bdf46f5786
r1238: resolve #589
2025-03-31 23:08:37 -04:00
Heng Li
f536b69b81
r1237: documented -x splice:sr
2025-03-30 21:47:47 -04:00
Heng Li
4b8b4418df
r1236: support paired-end short-read RNA-seq
2025-03-30 19:30:19 -04:00
Heng Li
ce30004e02
r1235: allow --splice and --frag at the same time
...
This seems to largely work, though the accuracy is reduced and no pairs are
proper. More investigation needed for practical uses.
2025-03-30 16:32:38 -04:00
Heng Li
54f8e5f7d6
r1234: added splice:sr for SE RNA-seq
2025-03-30 15:54:12 -04:00
Heng Li
fcb5d5e6eb
r1221: warn about file reading errors
...
Resolves #1229
2024-10-15 22:44:06 -04:00
Heng Li
3e1ab2951d
document --spsc
2024-10-12 23:27:46 -04:00
Heng Li
68179ed195
r1215: scoring apparently works
2024-10-12 22:29:27 -04:00
Heng Li
042c8d4d71
added --junc-pen; it does nothing for now
2024-10-11 21:27:25 -04:00
Heng Li
e4e1f7843b
backup for spsc
2024-10-10 22:43:50 -04:00
Heng Li
8170693de3
Release minimap2-2.28 (r1209)
2024-03-27 10:57:17 -04:00
Heng Li
e3d8c708ac
r1208: reverted RMQ gap coefficient
...
Such that minimap2 can give the same alignment in other modes
2024-03-27 08:48:10 -04:00
Heng Li
119bdc6029
r1207: reduced cap_kalloc from 1G to 500M
...
This reduces the peak memory.
2024-03-20 15:53:12 -04:00
Heng Li
89d4d219cd
r1206: enabled RMQ for lr:hqae
...
Also fixed a bug in determining inner_dist for RMQ. It should have no effect on
previous presets.
2024-03-20 15:29:54 -04:00
Heng Li
f51ff1abac
r1205: updated lr:hqae
2024-03-20 14:06:59 -04:00
Heng Li
c881b14ba5
r1203: added preset lr:hqae
2024-03-20 00:25:57 -04:00
Heng Li
f18dadb1c4
r1202: halved RMQ gap cost
2024-03-19 23:47:54 -04:00
