From owner-chemistry@ccl.net Mon Sep 10 01:15:01 2007 From: "Igor Pletnev igor.pletnev[A]gmail.com" To: CCL Subject: CCL: InChI version1.02beta; introducing InChIKey Message-Id: <-35110-070910011243-6077-/7BjtPAmPZ+aWRFAFqwNAQ%%server.ccl.net> X-Original-From: "Igor Pletnev" Content-Type: multipart/alternative; boundary="----=_Part_6775_31857658.1189401152890" Date: Sun, 9 Sep 2007 22:12:32 -0700 MIME-Version: 1.0 Sent to CCL by: "Igor Pletnev" [igor.pletnev^gmail.com] ------=_Part_6775_31857658.1189401152890 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline Dear Soaring Bear, let me clarify what is InChIKey and answer your questions. 1 > This change from a semi-readable smiles-like string to a string > which on casual appearance seems meaningless is a very > substantial change that belies the subtle version change from > 1.01 to 1.02. I have observed that the original InChI was just > coming into acceptance and wider use and now suddenly it is > being discarded. I'm sure there must have been very good First of all, please be aware that the original InChI is not discarded, absolutely. It remains a primary entity which may be parsed and back-converted to the structure. InChIKey is intended to be a _complementary_ key to InChI. The motivation to introduce InChIKey is to make web search and database indexing more easy. It is introduced for convenience and you may not use it at all, if your applications do not require this. As for appearance, an InChIKey string is as meaningless as any index key w.r.t. primary entity, e.g. as CAS Registry Number w.r.t. a structure, or passport number w.r.t. a person, or bar code w.r.t. a store item. 2 > is that calculated on a random theoretically diverse set of > structures or in the real world (of both nature and > combinatorial) where clustering occurs, and more (or less) > duplication may occur? Let me quote the release notes document included in the distribution, section "FAQ on InChIKey" . "Question 2: What is the collision resistance of the InChIKey? Has it been tested on real-life examples? Answer: InChIKey hash consists, internally, of 102 bits: 65 for the first block (molecular skeleton, or connectivity) and 37 for the second one (stereo/protonation/isotopic substitution isomers). This suggests that a theoretical - optimistic - estimate of the collision-resistance (corresponds to 50% probability of a single collision) is 6.1 billion molecular skeletons * 3.7*10**5 stereo/protonation/isotopic substitution isomers per each skeleton . = 2.2*10**15. Alternatively, for a collection of 1 billion different InChIKey entries, the estimated probability of an accidental collision of the first layers for a newly added structure is 2.7*10**(-9) % and for both layers is 2.0*10**(-20)%. The InChIKey was tested on databases of InChI strings created out of real and generated structures, derived from: Zinc (~4*10**6 entries, real structures, http://zinc.docking.org/), PubChem (~10*10**6, real, http://pubchem.ncbi.nlm.nih.gov/), GDB (~26*10**6, generated; courtesy of Prof. J.-L. Reymond, University of Berne, http://dcbwww.unibe.ch/groups/reymond/, private communication), FP42 (~42*10**6, custom-generated), OVERALL (~77*10**6, all of the above merged, duplicates excluded; real+generated). No hash collisions were observed in any of these databases. However, this does not imply that the collisions are impossible. In fact, due to the very essence of hash functions, collisions are unavoidable in sufficiently large collections. Some recent estimates of chemical space size for small molecules are in excess of 10**60, and for proteins it is 10**390 [Nature, 2004, 432(7019), Insight, pp. 823-865, http://www.nature.com/nature/insights/7019.html and refs. therein]. T. Fink et al. in 'Virtual Exploration of the Small-Molecule Chemical Universe below 160 Daltons', [Angew. Chem. Int. Ed. 2005, 44(10), pp. 1504-1508] quote an estimate of 10**18-10**200. Also, the estimates of hash collision probabilities given above are for an ideal hash and may not be valid in practice because of unknown yet properties of SHA-2 hash." Please note also that marking this release of v. 1.02 InChI software as 'beta' assumes, among other things, that a collision resistance of InChIKey will be subjected to independent testing by a community. Regards, Igor Pletnev --- Sent to CCL by: "Soaring Bear Ph.D." [soaringbear*yahoo.com] --- "steve heller srheller[#]nist.gov" wrote: > A new beta-release of the InChI software is now available > from the IUPAC web site ( www.iupac.org/inchi). > > The principal new features of this release are: > > (1) A fixed-length (25-character) condensed digital > representation of the Identifier to be known asInChIKey. ...... > Caffeine: InChI=1/C8H10N4O2/c1-10-4-9-6-5(10)7(13)12(3)8(14)11(6)2/h4H,1-3H3 > > InChIKey=RYYVLZVUVIJVGH-UHFFFAOYAW This change from a semi-readable smiles-like string to a string which on casual appearance seems meaningless is a very substantial change that belies the subtle version change from 1.01 to 1.02. I have observed that the original InChI was just coming into acceptance and wider use and now suddenly it is being discarded. I'm sure there must have been very good reasons for this change, but the magnitude of the change undermines confidence in the usage of it. What assurance is there that an equally major change isn't done in another 3 years? >....There is a finite, but very small probability of > finding two structures with the same InChIKey. For > duplication of only the first block of 14 characters this is > 1.3% in a thousand million, equivalent to a single collision > in one of 75 databases of one thousand million compounds > each. is that calculated on a random theoretically diverse set of structures or in the real world (of both nature and combinatorial) where clustering occurs, and more (or less) duplication may occur? Soaring Bear Ph.D. Pharmacology soaringbear*yahoo.com http://soaringbear.tripod.com/nature/weedsforneeds.html http://www.nlm.nih.gov/mesh/presentations/bear_2005_aug/index.htm author of http://HERBMED.org Dear Soaring Bear,
 
let me clarify what is InChIKey and answer your questions.
 
1
> This change from a semi-readable smiles-like string to a string
> which on casual appearance seems meaningless is a very
> substantial change that belies the subtle version change from
> 1.01 to 1.02.  I have observed that the original InChI was just
> coming into acceptance and wider use and now suddenly it is
> being discarded.  I'm sure there must have been very good
 
First of all, please be aware that the original InChI is not discarded, absolutely.
It remains a primary entity which may be parsed and back-converted to the structure.
 
InChIKey is intended to be a _complementary_ key to InChI.
The motivation to introduce InChIKey is to make web search and database indexing more easy.
It is introduced for convenience and you may not use it at all, if your applications do not require this.
 
As for appearance, an InChIKey string is as meaningless as any index key w.r.t. primary entity, e.g. as CAS Registry Number w.r.t. a structure, or passport number w.r.t. a person, or bar code w.r.t. a store item.

 
2
> is that calculated on a random theoretically diverse set of
> structures or in the real world (of both nature and
> combinatorial) where clustering occurs, and more (or less)
> duplication may occur?
 
Let me quote the release notes document included in the distribution, section "FAQ on InChIKey" .
 
"Question 2: What is the collision resistance of the InChIKey? Has it been tested on real-life examples?
 
Answer: InChIKey hash consists, internally, of 102 bits: 65 for the first block (molecular skeleton, or connectivity)
and 37 for the second one (stereo/protonation/isotopic substitution isomers).
 
This suggests that a theoretical - optimistic - estimate of the collision-resistance (corresponds to 50% probability of a single collision) is 6.1 billion molecular skeletons * 3.7*10**5 stereo/protonation/isotopic substitution isomers per each skeleton .
= 2.2*10**15.
 
Alternatively, for a collection of 1 billion different InChIKey entries, the estimated probability of an accidental collision of the first layers for a newly added structure is 2.7*10**(-9) % and for both layers is 2.0*10**(-20)%.
 
The InChIKey was tested on databases of InChI strings created out of real and generated structures, derived from:
 
Zinc (~4*10**6 entries, real structures, http://zinc.docking.org/),
PubChem (~10*10**6, real, http://pubchem.ncbi.nlm.nih.gov/),
GDB (~26*10**6, generated; courtesy of Prof. J.-L. Reymond, University of Berne, http://dcbwww.unibe.ch/groups/reymond/ , private communication),
FP42 (~42*10**6, custom-generated),
OVERALL (~77*10**6, all of the above merged, duplicates excluded; real+generated).
 
No hash collisions were observed in any of these databases.
 
However, this does not imply that the collisions are impossible.
In fact, due to the very essence of hash functions, collisions are unavoidable in sufficiently large collections.
Some recent estimates of chemical space size for small molecules are in excess of 10**60, and for proteins it is 10**390
[Nature, 2004, 432(7019), Insight, pp. 823-865, http://www.nature.com/nature/insights/7019.html and refs. therein]. T. Fink et al. in 'Virtual Exploration of the Small-Molecule Chemical Universe below 160 Daltons',
[Angew. Chem. Int. Ed. 2005, 44(10), pp. 1504-1508] quote an estimate of 10**18-10**200.
 
Also, the estimates of hash collision probabilities given above are for an ideal hash and may not be valid in practice because of unknown yet properties of SHA-2 hash."
 
Please note also that marking this release of v. 1.02 InChI software as 'beta' assumes, among other things, that a collision resistance of InChIKey will be subjected to independent testing by a community.
 
 
 
Regards,
Igor Pletnev
 

 

---
 
 
Sent to CCL by: "Soaring Bear Ph.D." [soaringbear*yahoo.com]


--- "steve heller srheller[#]nist.gov"
<owner-chemistry%%ccl.net> wrote:

> A new beta-release of the InChI software is now available
> from the IUPAC web site ( www.iupac.org/inchi).
>
> The principal new features of this release are:
>
> (1) A fixed-length (25-character) condensed digital
> representation of the Identifier to be known asInChIKey.
......
> Caffeine:
InChI=1/C8H10N4O2/c1-10-4-9-6-5(10)7(13)12(3)8(14)11(6)2/h4H,1-3H3
>
> InChIKey=RYYVLZVUVIJVGH-UHFFFAOYAW

This change from a semi-readable smiles-like string to a string
which on casual appearance seems meaningless is a very
substantial change that belies the subtle version change from
1.01 to 1.02.  I have observed that the original InChI was just
coming into acceptance and wider use and now suddenly it is
being discarded.  I'm sure there must have been very good
reasons for this change, but the magnitude of the change
undermines confidence in the usage of it.  What assurance is
there that an equally major change isn't done in another 3
years?



>....There is a finite, but very small probability of
> finding two structures with the same InChIKey. For
> duplication of only the first block of 14 characters this is
> 1.3% in a thousand million, equivalent to a single collision
> in one of 75 databases of one thousand million compounds
> each.

is that calculated on a random theoretically diverse set of
structures or in the real world (of both nature and
combinatorial) where clustering occurs, and more (or less)
duplication may occur?




Soaring Bear Ph.D. Pharmacology  soaringbear* yahoo.com
http://soaringbear.tripod.com/nature/weedsforneeds.html
http://www.nlm.nih.gov/mesh/presentations/bear_2005_aug/index.htm
author of http://HERBMED.org



E-mail to subscribers: CHEMISTRY[]ccl.net or use:
     http://www.ccl.net/cgi-bin/ccl/send_ccl_message

E-mail to administrators: CHEMISTRY-REQUEST[]ccl.net or use
     http://www.ccl.net/cgi-bin/ccl/send_ccl_message

Subscribe/Unsubscribe:
     http://www.ccl.net/chemistry/sub_unsub.shtml

Before posting, check wait time at: http://www.ccl.net

Job: http://www.ccl.net/jobs
Conferences: http://server.ccl.net/chemistry/announcements/conferences/

Search Messages: http://www.ccl.net/htdig  (login: ccl, Password: search)
     http://www.ccl.net/spammers.txt

RTFI: http://www.ccl.net/chemistry/aboutccl/instructions/



 

------=_Part_6775_31857658.1189401152890-- From owner-chemistry@ccl.net Mon Sep 10 04:26:01 2007 From: "Sina T reli sinatureli()gmail.com" To: CCL Subject: CCL: Spartan: Bonds in GUI Message-Id: <-35111-070910040801-6690-rcnfnP8Q9IyCn7cjmJ7ZKA/./server.ccl.net> X-Original-From: "Sina T reli" Date: Mon, 10 Sep 2007 04:07:57 -0400 Sent to CCL by: "Sina T reli" [sinatureli~!~gmail.com] Hello, I am using spartan to do some studies on chlorophylls. I am using samples (chlorophyll a's) from a 2A resolution pdb file. In the one I am using in the windows there seems to be bonds between the four nitrogen and the MG. In some representations there are two bonds and in some pdb files there are none. I would think that it would do no difference whether if I added bonds or deleted them from the windows. But when I try to coordinate (first using energy minimize with only the solvent molecule able to move) the MG atom with a water or acetone, if the mg is bonded to the nitrogens, the polar side of the solvent molecule tends to turn away from the molecule. Where as if I delete the bonds manually, the solvent molecule is coordinated with MG atom. Also spartan makes some errors assuming that some of the double bonded oxygens around the chlorophyll is single bonded (which I refine manually). Also in constrast to classical representation of chlorophylls, the ring is almost fully conjugated (which is fine anyway). In short should I trust spartan or should I refine the molecula manually because in fact adding bonds between MG and its coordinators change the outcome of the results (even if slightly...). Or could you suggest me a program that finely refines such errors. Thank you very much.. From owner-chemistry@ccl.net Mon Sep 10 05:36:01 2007 From: "Maxim Kholin maxim.kholin|a|q-pharm.com" To: CCL Subject: CCL: query reagrding docking through Molecular Dynamics Message-Id: <-35112-070910021450-32022-aBQdCsXLxfYbbUO7dtoXLQ~~server.ccl.net> X-Original-From: "Maxim Kholin" Content-Type: multipart/alternative; boundary="----=_NextPart_000_0257_01C7F393.5355C320" Date: Mon, 10 Sep 2007 10:14:08 +0400 MIME-Version: 1.0 Sent to CCL by: "Maxim Kholin" [maxim.kholin() q-pharm.com] This is a multi-part message in MIME format. ------=_NextPart_000_0257_01C7F393.5355C320 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, Quantum docking software allows that, it docks on fully flexible = protein. Since it is pure physics based approach an accuracy of IC50 = predictions are also good for absolutely new chemical classes. You can change sequences in the protein, make molecular dynamics and = dock it again, thus researching changes in the activity after mutations. The error of prediction is 1 pkd for all chemical classes, that is very = good. Take free demo at www.q-pharm.com. We can also make in-house ic50 predictions for protein-protein complex = as a service. Feel free to ask questions, Maxim Kholin ----- Original Message -----=20 From: Kalyan chaitanya kalyanpulipaka^-^gmail.com=20 To: Kholin, Maxim N =20 Sent: Saturday, September 08, 2007 8:39 PM Subject: CCL: query reagrding docking through Molecular Dynamics Dear all, Presently am working on Molecular docking, both protein-ligand and = protein-protein docking now iam intrested in tlooking at conformational = changes of receptor while docking, which is possible to through = Molecular Dynamics (as per my knowledge), but i don't know as to what = softwares are available for the same purpose and also please let me know = if any one has any tutorrial for the same.=20 Thank you in advance, Regards, Kalyan. ------=_NextPart_000_0257_01C7F393.5355C320 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Hi,
Quantum docking software allows that, = it docks on=20 fully flexible protein. Since it is pure physics based approach an = accuracy of=20 IC50     predictions are also good for absolutely new = chemical=20 classes.
You can change sequences in the = protein, make=20 molecular dynamics and dock it again, thus researching changes in the = activity=20 after mutations.
The error of prediction is 1 pkd for = all chemical=20 classes, that is very good.
Take free demo at www.q-pharm.com.
We can also make in-house ic50 = predictions for=20 protein-protein complex as a service.
Feel free to ask = questions,
Maxim Kholin
 
 
----- Original Message -----
From:=20 Kalyan=20 chaitanya kalyanpulipaka^-^gmail.com
Sent: Saturday, September 08, = 2007 8:39=20 PM
Subject: CCL: query reagrding = docking=20 through Molecular Dynamics

Dear all,

Presently am working on Molecular docking, both protein-ligand and=20 protein-protein docking now iam intrested in tlooking at = conformational=20 changes of receptor while docking, which is possible to through = Molecular=20 Dynamics (as per my knowledge), but i don't know as to what softwares = are=20 available for the same purpose and also please let me know if any one = has any=20 tutorrial for the same.

Thank you in advance,

Regards,

Kalyan.

------=_NextPart_000_0257_01C7F393.5355C320-- From owner-chemistry@ccl.net Mon Sep 10 05:37:01 2007 From: "Jens Spanget-Larsen spanget!=!ruc.dk" To: CCL Subject: CCL: IR spectrum by MS Modeling (Intensity kilometer per mole) Message-Id: <-35113-070910051017-28875-SgOtoplVBBcb3k/rwMjEJg---server.ccl.net> X-Original-From: Jens Spanget-Larsen Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 10 Sep 2007 10:11:20 +0200 MIME-Version: 1.0 Sent to CCL by: Jens Spanget-Larsen [spanget,+,ruc.dk] Dear Jennie! The relation between 'Absorbance' (dimensionless) and 'IR intensity' in km/mol is described on my web-page: http://www.ruc.dk/~spanget/IR_intensity.doc YOurs, Jens >--< ------------------------------------------------------ JENS SPANGET-LARSEN Office: +45 4674 2710 Dept. of Science (18.1) Fax: +45 4674 3011 Roskilde University Mobile: +45 2320 6246 P.O.Box 260 E-Mail: spanget[A]ruc.dk DK-4000 Roskilde, Denmark http://www.ruc.dk/~spanget ------------------------------------------------------ Jennie Gao sci11075^gmail.com wrote: > Sent to CCL by: "Jennie Gao" [sci11075|*|gmail.com] Dear All, > > I came across some difficulties regarding the IR spectrum obtained by > MS Modeling and would like to seek for some opinion from you. > > Recently i had calculated the IR spectrum by MS Modeling. I realized > that the intensity was in kilometer permole (kmmol-1). How can i > compare it with the experimental result where the intensity is in > Absorbance (A)? > > Thank you so much for your assistance. > > Best Regards Jennie From owner-chemistry@ccl.net Mon Sep 10 06:43:01 2007 From: "chupvl chupvl++gmail.com" To: CCL Subject: CCL: query reagrding docking through Molecular Dynamics Message-Id: <-35114-070910045335-19968-l8xjCaY2QrhjHqTfxaEEYg(a)server.ccl.net> X-Original-From: chupvl Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 10 Sep 2007 11:56:16 +0400 MIME-Version: 1.0 Sent to CCL by: chupvl [chupvl-#-gmail.com] Dear Kalyan, The well known molecular dynamics (MD) software packages are - AMBER - NAMD - http://www.ks.uiuc.edu/Research/namd/ - Gromacs and lot of other useful MD packages you will know just wikiing the "molecular dynamics" term http://en.wikipedia.org/wiki/Molecular_dynamics Sincerely yours, Chupakhin Vladimir Kalyan chaitanya kalyanpulipaka^-^gmail.com wrote: > > Dear all, > > Presently am working on Molecular docking, both protein-ligand and > protein-protein docking now iam intrested in tlooking at > conformational changes of receptor while docking, which is possible to > through Molecular Dynamics (as per my knowledge), but i don't know as > to what softwares are available for the same purpose and also please > let me know if any one has any tutorrial for the same. > > Thank you in advance, > > Regards, > > Kalyan. > From owner-chemistry@ccl.net Mon Sep 10 07:16:01 2007 From: "Herbert Fruchtl herbert.fruchtl%a%st-andrews.ac.uk" To: CCL Subject: CCL: scan Message-Id: <-35115-070910064145-19151-6wh0IGZ9or7AyzimESmmTg a server.ccl.net> X-Original-From: Herbert Fruchtl Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 10 Sep 2007 11:40:52 +0100 MIME-Version: 1.0 Sent to CCL by: Herbert Fruchtl [herbert.fruchtl-*-st-andrews.ac.uk] A few points, without looking at the details of the system you are studying: SCF (including DFT) finds a stationary point (hopefully a minimum) in the orbital space. This is not guaranteed to be the global minimum. It's just the minimum you find when you start searching from your initial guess. In a scan, the solution from the previous geometry is taken as the initial guess. This is good as long as the optimal electron configuration is still the same as at the previous geometry. If at a given geometry a new, different, electron configuration would give a lower energy, but the old one is still a local minimum, there is a good chance that you find the old, higher one. At some point the local minimum disappears, and the SCF finds the lower configuration. This means a discontinuity in the PES. The points just before this drop are almost certainly invalid. If you do the scan in the opposite direction, you will get rid of some of the problems. Your transition state won't be very reliable though, because the electron configuration there is likely of multi-reference character, and simple B3LYP won't describe this. As for your convergence problems, SCF=QC or XQC often (but not always) helps, but makes things slower. Guess=always will prevent using the previous solution as guess, but will also make things slower, and of course still doesn't guarantee that you find the global minimum. As an aside: I am always wary of using different quality basis sets for similar atoms. 6-31g* on N and O, but sto-3g on C sounds dodgy. Use at least 6-31g for C and H. HTH, Herbert hadi behzadi hadi.behzadi**gmail.com wrote: > Sent to CCL by: "hadi behzadi" [hadi.behzadi*_*gmail.com] > Dear all, > > I am trying to do a PES scan for Fe-O distance between 1.75-8 A in heme with one and two water as a axial ligand. I have two problem in this relation as follow: > > 1- For heme with one water the calculation give me error in 4.1 A because of failure to convergence, the energy profile is good befor this distance.I use SCF(maxcycle=5000) and the B3lyp method with the follow basis sets: > n,o 6-31g* > c,h sto-3g > fe lanl2dz > > 2- For heme with two water as a axial ligand,one water is fixed, the energy profile is not good due to the break in energy profile in 2.35 and 2.45 A. Here is the obtained energy and used basis sets: > > n,o,c,h 3-21g* > fe lanl2dz > Distance Energy(A.U) > 1.75 -2392.47636390 > 1.85 -2392.48624467 > 1.95 -2392.48987980 > 2.05 -2392.49028130 > 2.15 -2392.48885147 > 2.25 -2392.48631707 > 2.35 -2392.48316683 > 2.45 -2392.58475917 > 2.55 -2392.58231643 > 2.65 -2392.57967485 > 2.75 -2392.57698884 > 2.85 -2392.57434438 > 2.95 -2392.57179832 > 3.05 -2392.56938208 > 3.15 -2392.56710231 > > I really don t underestand what is the problem and any Help would be appreciated. > Regards> > > -- Herbert Fruchtl EaStCHEM Fellow School of Chemistry University of St Andrews From owner-chemistry@ccl.net Mon Sep 10 08:58:00 2007 From: "hans himmler contact . hjhimmler.de" To: CCL Subject: CCL: activity prediction software Message-Id: <-35116-070910063912-18585-BtDTh5VEM6DS1zO811ALQw[#]server.ccl.net> X-Original-From: "hans himmler" Date: Mon, 10 Sep 2007 06:39:09 -0400 Sent to CCL by: "hans himmler" [contact||hjhimmler.de] I want to do some work on clustering molecular structures based on predicted biological activities. In that context I am looking for free or low cost programs that predicts activities. The program (or package) must run on windows, must provide an API or a command line interface and must support structure input via SDFiles (or smiles). Thanks for feedback in advance. From owner-chemistry@ccl.net Mon Sep 10 09:51:00 2007 From: "Debellis Anthony CE US anthony.debellis:_:cibasc.com" To: CCL Subject: CCL: Spartan: Bonds in GUI Message-Id: <-35117-070910094311-23280-4tjI+jegi7Mpd9mm7q3Taw!A!server.ccl.net> X-Original-From: "Debellis Anthony CE US" Content-class: urn:content-classes:message Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset="US-ASCII" Date: Mon, 10 Sep 2007 09:36:56 -0400 MIME-Version: 1.0 Sent to CCL by: "Debellis Anthony CE US" [anthony.debellis[-]cibasc.com] Hello, The bonds in Spartan have meaning only in the molecular mechanics code (energy minimize in the builder/GUI). The GUI should be used to provide only a reasonable starting point for input to the other (ab initio, DFT, or semiemp) modules. Anthony -----Original Message----- > From: owner-chemistry++ccl.net [mailto:owner-chemistry++ccl.net] Sent: Monday, September 10, 2007 4:08 AM To: Debellis Anthony CE US Subject: CCL: Spartan: Bonds in GUI Sent to CCL by: "Sina T reli" [sinatureli~!~gmail.com] Hello, I am using spartan to do some studies on chlorophylls. I am using samples (chlorophyll a's) from a 2A resolution pdb file. In the one I am using in the windows there seems to be bonds between the four nitrogen and the MG. In some representations there are two bonds and in some pdb files there are none. I would think that it would do no difference whether if I added bonds or deleted them from the windows. But when I try to coordinate (first using energy minimize with only the solvent molecule able to move) the MG atom with a water or acetone, if the mg is bonded to the nitrogens, the polar side of the solvent molecule tends to turn away from the molecule. Where as if I delete the bonds manually, the solvent molecule is coordinated with MG atom. Also spartan makes some errors assuming that some of the double bonded oxygens around the chlorophyll is single bonded (which I refine manually). Also in constrast to classical representation of chlorophylls, the ring is almost fully conjugated (which is fine anyway). In short should I trust spartan or should I refine the molecula manually because in fact adding bonds between MG and its coordinators change the outcome of the results (even if slightly...). Or could you suggest me a program that finely refines such errors. Thank you very much..http://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt From owner-chemistry@ccl.net Mon Sep 10 10:26:01 2007 From: "chupvl chupvl[]gmail.com" To: CCL Subject: CCL: activity prediction software Message-Id: <-35118-070910094839-25946-9P5A8PsTgh2xd7Zn/liw3Q=-=server.ccl.net> X-Original-From: chupvl Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 10 Sep 2007 17:45:54 +0400 MIME-Version: 1.0 Sent to CCL by: chupvl [chupvl^gmail.com] Hello! It's not clear You want to cluster structures (dataset)? >> Solution is Maximal common substructure (subgraph) algorithm used by CDK, Joelib and any others software. or You want to cluster structures based only on activities? I think there is know free software...but you can use commercial SARvision or You want to build QSAR models based on clustering? Recursive partitioning is now used only by commercial software (as I know). Maybe (SCAM (Statistical Classification of Activities of Molecules) will be solution for you but I don't know available and "live" package. Sincerely yours, Chupakhin Vladimir MSU, Dpt. of Chemistry, Moscow, Russia Maximal common substructure (subgraph) hans himmler contact . hjhimmler.de wrote: > Sent to CCL by: "hans himmler" [contact||hjhimmler.de] > I want to do some work on clustering molecular structures based on predicted biological activities. > In that context I am looking for free or low cost programs that predicts activities. > The program (or package) must run on windows, must provide an API or a command line interface and must support structure input via SDFiles (or smiles). > Thanks for feedback in advance.> > > > > From owner-chemistry@ccl.net Mon Sep 10 11:01:00 2007 From: "Igor Filippov Contr igorf||helix.nih.gov" To: CCL Subject: CCL: activity prediction software Message-Id: <-35119-070910095417-28090-1NZwHioIRXlSkqcjZVBJMQ+*+server.ccl.net> X-Original-From: "Igor Filippov [Contr]" Content-Transfer-Encoding: 7bit Content-Type: text/plain Date: Mon, 10 Sep 2007 09:54:08 -0400 Mime-Version: 1.0 Sent to CCL by: "Igor Filippov [Contr]" [igorf|,|helix.nih.gov] PASS (prediction of activity spectra) seems to be exactly what you need: http://www.ibmc.msk.ru/sections_en/Sect4_ProdServ/item3_soft/page.shtml It runs on windows and has both GUI and a command-line interface. It takes SD files as input. Igor On Mon, 2007-09-10 at 06:39 -0400, hans himmler contact . hjhimmler.de wrote: > Sent to CCL by: "hans himmler" [contact||hjhimmler.de] > I want to do some work on clustering molecular structures based on predicted biological activities. > In that context I am looking for free or low cost programs that predicts activities. > The program (or package) must run on windows, must provide an API or a command line interface and must support structure input via SDFiles (or smiles). > Thanks for feedback in advance.-- Igor Filippov [Contr] From owner-chemistry@ccl.net Mon Sep 10 11:47:00 2007 From: "Soaring Bear Ph.D. soaringbear^^yahoo.com" To: CCL Subject: CCL: activity prediction software Message-Id: <-35120-070910111408-21627-Kr20jGp+aWmT3wIi3h8WRQ|-|server.ccl.net> X-Original-From: "Soaring Bear Ph.D." Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=iso-8859-1 Date: Mon, 10 Sep 2007 08:13:59 -0700 (PDT) MIME-Version: 1.0 Sent to CCL by: "Soaring Bear Ph.D." [soaringbear]*[yahoo.com] --- "hans himmler contact . hjhimmler.de" wrote: > I want to do some work on clustering molecular structures > based on predicted biological activities. That is a very difficult thing to do without some kind of activity data to start with and even then such QSAR is difficult to project beyond the extent of the data. > In that context I am looking for free or low cost programs > that predicts activities. Well, there is PASS http://195.178.207.233/PASS/BAS.html Also, this type of thing is being pursued by pubchem in the properties and activities that are being added to that chemical database http://www.ncbi.nlm.nih.gov/sites/entrez?db=pccompound Soaring Bear Ph.D. Pharmacology soaringbear at yahoo.com http://geocities.com/soaringbear/nature/weedsforneeds.html http://www.nlm.nih.gov/mesh/presentations/bear_2005_aug/index.htm author of http://HERBMED.org From owner-chemistry@ccl.net Mon Sep 10 12:20:01 2007 From: "Akef Taha Afaneh akef_afnh()yahoo.com" To: CCL Subject: CCL: scan Message-Id: <-35121-070910101812-10304-6u5dFZE0AXeoDNfgL2Ieew%a%server.ccl.net> X-Original-From: "Akef Taha Afaneh" Date: Mon, 10 Sep 2007 10:18:08 -0400 Sent to CCL by: "Akef Taha Afaneh" [akef_afnh]![yahoo.com] I think you must use at least 6-31G* basis set for the C and H atoms.