Tailored solutions for Bioinformatics

01. Service Level

E-health Web Portal

Challenge:

Healthcare requires reliable software since even minor system errors can lead to critical consequences. It is essential to pay additional attention to the automation of some routine processes that are carried out manually.

Azati developed a web portal, which provides medical services online, increasing work efficiency and reducing bureaucracy.

Solution:

To solve the problems with bureaucracy and slow end-user feedback, the engineers developed a web application that provides some medical services online. However, our team did not participate in the initial portal development.

Azati modernized the existing version, transfer the software to the cutting-edge tools, and enhanced the already existing functionality to provide more facilities.

02. Service Level

BLAST Algorithm Enhancement

Challenge:

While working with genomic sequences and conducting biological research, scientists rely on massive databases. The navigation and searching for the necessary information in such databases become quite tricky, using the conventional searching techniques.

To solve this problem, the customer implemented the BLAST algorithm. However, this algorithm had flaws and was not very accurate.

Solution:

To optimize the BLAST algorithm, our engineers analyzed the source code of the solution. Engineers tuned existing algorithm parameters and internal weights, as well as developed additional search filters to accelerate the search process and make it more accurate.

After some improvements, the algorithm provides more accurate and precise results. All these changes made the search engine more accurate and flexible to customer needs.

Service Level

01. Tailored Solutions

80 Fold Software Performance Improvement

Challenge:

During the DNA sequencing process, the customer processes millions of records per pay. The software built for this purpose was not designed to handle such amounts of information. Specialists spent hours waiting for results.

Our objective was to increase the productivity of the algorithms and boost efficiency within the shortest possible time.

Solution:

After the initial analysis, the engineers discovered the data processing algorithms were rather appropriate. Still, its implementations in Perl were poor and significantly stretched the time to process the presented sequence data.

Azati decided to rewrite the source code in C++, and its programming paradigms positively affected the processing duration.

02. Tailored Solutions

Search Engine for Multiple Genetics Sequence Search

Challenge:

While working with medical data such as fusion/chimeric constructs, recombinant plasmid constructs scientists rely on searching methods based on biological sequences in various sets of nucleotides and peptides.

The customer needed a reliable and accurate tool to search for multiple nucleotide queries across massive nucleotide databases.

Solution:

The Azati team has designed, developed and released a universal multi-sequence search tool (MSS) and simultaneous multi-CDR search, which can process up to six query sequences simultaneously.

Our engineers have implemented the Smith-Waterman algorithm, which fully satisfies the requirements and brings a higher degree of sophistication to the researchers.

03. Tailored Solutions

Automating IP Processes

Challenge:

Bioinformatics requires precision in all processes since all the details play an important role in genetic research. It’s always been crucial for a researcher to make sure that he receives actual results in time.

As the web portal automatically updates data every 24 hours, it is challenging to keep an eye on the updates of patent documents.

Solution:

Azati created a small module that implements a monitoring feature that helps users independently configure updates of their preferred searches. The team also added several additional parameters to internal search algorithms to simplify monitoring integration.

As a result, users can see the up-to-date search results staying relieved from manual and constant routine search.

04. Tailored Solutions

Genetic Analysis Tool

Challenge:

The accuracy and relevance of biological research are crucial in achieving the goal. The volume of legal scientific information is growing exponentially every year, which significantly complicates the timely information update.

Researchers were forced to check several databases manually, spending lots of time without guaranteeing the result correctness.

Solution:

The Azati team successfully tackled the challenge by developing and maintaining the online IP Research Portal with easy-to-use, readily accessible content, search, analysis, and reporting tools.

The portal provides an opportunity to perform searches by the sequences query using such algorithms as BLAST, Smith-Waterman, Multiple Sequence Search algorithm and MOTIF.

Tailored Solutions

01. Research & Development

The Smith-Waterman Algorithm Improvement

Challenge:

Many scientists widely use the Smith-Waterman algorithm during their DNA research. The user has to spend a lot of time to receive accurate results, as it takes a lot of time for the algorithm to complete its work.

The customer asked Azati to optimize the existing algorithm and reduce the data processing duration.

Solution:

Azati has accelerated the algorithm by using cloud and GPU computing in combination with improvements in the specialized genetic alignment technology.

Azati enriched the workflow with cloud computing to handle big data processing and provide the final result as fast as possible. Thus, the search time reduced by 30-50 times.

02. Research & Development

Semantic Search Engine

Challenge:

Huge sets of scientific data have structural complexity, and a large number of interrelated terms sophisticate and delay data processing. The scientists engaged in blood sample manipulating had to define tags manually.

Bioinformatics companies need powerful search engines that can help to deal with the internal inventory search.

Solution:

Our specialists built a solution consisting of three separate modules located in the cloud: Query Analysis module, Search Engine module, and User Interface API.

Two modules are powered by machine learning to process data in a natural language and to compare extracted entities with synonyms for the fast and accurate search.