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:20250822T115806Z
LOCATION:Room 6.0D13
DTSTART;TZID=Europe/Stockholm:20250616T173000
DTEND;TZID=Europe/Stockholm:20250616T180000
UID:submissions.pasc-conference.org_PASC25_sess169_pap123@linklings.com
SUMMARY:Accelerated CNN-based Scans for Traces of Positive Selection
DESCRIPTION:Sjoerd van den Belt and Nikolaos Alachiotis (University of Twe
 nte)\n\nPositive natural selection is the driving force that enables speci
 es<br />to survive and reproduce in their environment. Localizing traces o
 f positive selection has practical applications in studying virus evolutio
 n and designing more effective drug treatments. State-of-the-art methods f
 or the detection of positive selection combine Convolutional Neural Networ
 ks (CNN) with sliding-window algorithms to scan genomic sequences with hig
 h precision, but require prohibitively long execution times to process who
 le genomes with fine-grained resolution. We present an FPGA-accelerated sy
 stem for efficiently scanning whole genomes with high granularity, impleme
 nting a quantized version of FAST-NN, a CNN that has been designed through
  a hardware-aware neural architecture search. FAST-NN employs a compact re
 presentation of genomic data as features, which eliminates potential I/O b
 ottlenecks in hardware. Our accelerator architecture consists of a dedicat
 ed stage for each CNN layer in a pipelined datapath that integrates a spec
 ialized buffer design; this enables data reuse between overlapping sliding
  windows by leveraging the dilated convolutions in FAST-NN. A design point
  implemented onto an Alveo U250 accelerator card achieves comparable accur
 acy to FAST-NN, with a maximum reduction of only 2.2% due to quantization,
  while producing a classification outcome in each clock cycle at a frequen
 cy of 100MHz. Scanning the entire human genome (excluding the sex chromoso
 mes), we observed between 19.51× and 28.61× faster processing than a PyTor
 ch implementation on a 16-core CPU, and between 1.22× and 2.89× faster pro
 cessing than a high-end GPU. The architecture is adaptable to other domain
 s where CNNs are deployed in sliding-window algorithms for large-scale dat
 a processing.\n\nDomain: Engineering, Computational Methods and Applied Ma
 thematics\n\nSession Chair: Nina Mujkanovic (ETH Zurich / CSCS)\n\n
END:VEVENT
END:VCALENDAR