BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20190719T085745Z
LOCATION:HG E 1.2
DTSTART;TZID=Europe/Stockholm:20190614T120000
DTEND;TZID=Europe/Stockholm:20190614T123000
UID:submissions.pasc-conference.org_PASC19_sess127_msa305@linklings.com
SUMMARY:De Novo Drug Design with Artificial Intelligence
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Emerg
 ing Application Domains, Chemistry and Materials, Life Sciences\n\nDe Novo
  Drug Design with Artificial Intelligence\n\nSchneider\n\nWe have implemen
 ted and challenged 'chemistry-savvy' deep learning models in prospective m
 olecular design projects that aimed to obtain synthetically easily accessi
 ble new chemical entities. We will present several prospective application
 s and discuss opportunities and current limitations of theses computationa
 l models. In a first study, recurrent networks were trained with structure
 s of known synthetically accessible, druglike compounds. By applying trans
 fer learning, the learned feature distributions were biased towards certai
 n pharmacologically desired endpoints. The computationally generated de no
 vo designs were subsequently prioritized, chemically synthesized and bioch
 emically tested for the predicted activities with high success rates. In t
 he second study, we developed a novel virtual synthetic assembly method th
 at combines a rule-based approach with a neural network trained on success
 ful synthetic routes described in chemical patent literature. This unique 
 combination enabled a balance between ligand-similarity based generation o
 f innovative compounds by scaffold hopping and forward-synthetic feasibili
 ty of the designs. In a prospective proof-of-concept application, the soft
 ware generated sets of de novo designs for four approved drugs that were i
 n agreement with the desired structural and physicochemical properties. Se
 lected computer-generated compounds were successfully synthesized in accor
 dance with the synthetic route proposed by this method.
END:VEVENT
END:VCALENDAR