Codondex Biohunt Grants

Codondex has developed an algorithmic investigation tool to analyse genetic sequences at a resolution of an arbitrary number of nucleotides. The tool allows us to obtain an easy way to observe the fine repetitive details of small subsequences contained within a gene.

Using this tool, we compute various metrics for each subsequence including Inclusiveness, which measures the total number of occurrences of every computed subsequence found within the subsequence of interest.

Our primary interest is in intronic, non-coding DNA in multi-transcript genes. In these systems we can create a transcript list, which we call the Vector, that is sorted by Codondex i-Score (a metric which looks at Inclusiveness scaled by the length of the subsequence, to better account for intrinsic probability of finding smaller subsequences within progressively longer ones). Using this we look at the way the order of this vector changes from subsequence to subsequence. Large changes in these vectors then prompts us to tag them for further investigation as it represents large deviation from transcript similarity, with this subsequence being labelled a Key Sequence.

Codondex is proposing 3 grants for open problems to aid in our journey towards a more biologically useful platform. These 3 problems span statistical analysis, data acquisition and biological relevance of various aspects that are integral to our platform.

Applicants should sign up and email adam@codondex.com with proposals or questions pertaining to the grants.

Task 1: Exploration of statistical methods for analysis of i-Score reports.

Problems:

• Large dataset, which by the nature of the algorithm produces a large amount of redundant data
• Majority of the set is uninteresting in terms of vector changes
• Changes in one offset may be directly related to changes in another offset if there is overlapping sequence start/end
• Weak metrics for observing interesting features and future analysis.

Criteria:

• Formal statistical analysis of the dataset
  • Unsupervised/supervised clustering
  • Principal Component Analysis
  • Machine learning
• Algorithm for reduction of dataset by removing inconsequential datapoints/removing redundant datapoints.
• Formulation of a metric or metrics that can flatten the set (scales inversely with length among other characteristics) and identify offsets of particular importance or containing interesting features with respect to the rest of the set.

Judgement:

An internal panel as well as a consultant to judge the validity of the results will determine if the work is of sufficient quality.

Grant funds:

US$10,000. 25% upon acceptance of the grant, 75% upon completion to a sufficient standard.

Timeline:

Approximately 8 weeks.

Task 2: Importation and collation of patient TP53 genes containing indels and/or SNPs on a mass scale.

Problems:

• Lack of an efficient way of collecting the large amount of available publicly available patient genome data
• Lack of an efficient way to collate the data into a format efficient for our purposes

Criteria:

• A way to address the 2 problems efficiently along with pairwise alignments to the current human genome assembly for each of the patient TP53 genes.

Judgement:

An internal panel will determine if the criteria have been met.

Grant funds:

US$5,000. 20% upon acceptance of the grant, 80% upon completion to a sufficient standard.

Timeline:

Approximately 4 weeks.

Task 3: Investigation of biological relevance i-Score report analysis.

Problems:

• Finding biological significance from big data in bioinformatics is a great challenge for the field.
• We believe our algorithm and metrics are identifying intronic sequences of biological importance but lack the expertise to efficiently compare our data to existing databases.

Criteria:

• From a provided key sequences report, search databases for the sequences with known biological significance.
• For each hit provide a short summary of its purported function.

Bonus:

• Design an experiment to test the in vitro or in vivo effects of these biologically relevant sequences.

Judgement:

An internal panel will determine if the criteria have been met.

Grant funds:

US$5,000. 20% upon acceptance of the grant, 80% upon completion. Submission of an experimental design that we believe to be worth conducting will be rewarded with an extra US$1000. If you have the ability to perform the submitted experimental design, we will open discussions for funding it.