SpliceCraft v1.2.57

Bug fixes

• Enzymes sold under another name couldn't be found. The same restriction
  enzyme is sold under different names by different suppliers — Eco31I is
  BsaI, LguI is SapI, AarI is PaqCI, BveI is
  BspMI, TspMI is XmaI — and asking for one SpliceCraft didn't
  list came back as an unrecognised name, on an enzyme it knows perfectly
  well. Thirty-nine of the names people actually order now resolve, and your
  sites come back labelled with the name you asked for. Where two enzymes
  read the same bases but cut in different places they stay separate, since
  they leave different ends: Acc65I (G^GTACC) has been added alongside
  KpnI as its own enzyme rather than treated as another word for it.
• Two enzymes that cut the same site differently reported as one. XmaI
  and SmaI both recognise CCCGGG but cut it in different places — one
  leaves a sticky end, the other is blunt. The plasmid map deliberately draws
  a single bar there so sites don't stack up, and the scripting API was
  reading its answer off that drawing, so asking about both returned only
  one. Now it reports both, with each enzyme's real cut position. Same for
  Acc65I/KpnI, ApaI/PspOMI, NheI/BmtI, KasI/NarI, AatII/ZraI and
  SacI/Eco53kI. The map is unchanged.
• Golden Gate simulation blamed the parts for a problem with the vector.
  If the destination plasmid carried a stray site for the assembly enzyme it
  got cut into more than two pieces, which is fine — a real one-pot reaction
  puts all of them back around the inserts — but the simulator assumed the
  backbone was a single piece and reported "the overhangs don't chain, check
  that adjacent parts share a 4 nt overhang". The parts were never the
  problem. Multi-piece backbones now assemble, and when a reaction genuinely
  can't close, the message shows its working: how many times the vector was
  cut and where, and every fragment's two ends, so the one with no partner is
  visible. If more than one arrangement could form, you're told rather than
  handed one of them — and if many fragments share an overhang, the search is
  cut off with an honest "unknown" rather than grinding through an
  astronomical number of orderings or claiming the design can't be built.
• Primer hairpin and self-dimer energies were a thousand times too large.
  They came back as -738 and -4993 where an oligo's real values are
  single digits — reported in calories where the whole field, and every rule
  of thumb you'd compare them against, uses kilocalories. They're now in
  kcal/mol, each result says so, and the figure quoted for a primer pair is
  the worse of the two rather than the better one. If the calculation can't
  run at all you get a blank and a warning instead of a zero, which used to
  read as a perfect score.
• Site cures could stretch a run of identical bases. Removing a
  restriction site sometimes has two equally silent single-base fixes, and
  the one SpliceCraft picked could butt onto an existing run — curing
  GGAAAAA GAAGAC produced GGAAAAAAAAGAC, eight A's in a row, where the
  other choice left the longest run at five. Long runs are hard to synthesise
  and hard to sequence, and when the site sits in a repeated element (a
  tandem 2×35S enhancer, say) the same bad choice lands once per copy. Cures
  now prefer the option that doesn't lengthen a run, ranked above the
  codon-usage preference. Every plan reports max_homopolymer_run, and if no
  silent alternative avoided lengthening one, it says so instead of leaving
  you to check.
• A misspelt enzyme made scrub-plasmid report a clean plasmid. Names it
  didn't recognise were quietly skipped, so a typo (or an enzyme entered in
  the wrong capitalisation, or under a supplier's name) scrubbed nothing and
  reported success with no sites removed — which reads as "already clean" on
  a plasmid that isn't. Unrecognised names are now a clear error, and
  bsai / Eco31I work like BsaI.
• A feature crossing the start of the plasmid came out of a clone in two
  pieces. Carrying annotations through a restriction clone splits every
  origin-crossing feature in half while the plasmid is cut open — necessary —
  but nothing put them back together afterwards. A T-DNA border or marker
  nowhere near the cloning site arrived as two adjacent features with the
  same name, which reads as a broken border. They're rejoined now, unless the
  cloning really did separate them.