Enhance Your Search Sensitivity with AI: Off-Target Effects of BRAF

Targeting a specific gene can inadvertently affect other targets and biological pathways,¹ leading to unintended consequences. Understanding the off-target gene effects can help to predict and mitigate potential adverse reactions that could compromise the safety and efficacy of drugs.² This blog explores how advanced tools like Causaly and traditional databases like PubMed approach the challenge of identifying relevant literature on the off-target effects of genes. Through this comparison, we underscore the importance of leveraging AI in biomedical research, paving the way for safer and more effective therapeutic solutions.

We explored the off-target effects of BRAF using both Causaly and PubMed. Using Causaly’s AI, 109 total articles relevant to the search were identified, compared to 92 papers returned when performing the same search in PubMed.

Of the papers found in PubMed, 53% were also retrieved by Causaly (49 of 92). The remaining papers identified in PubMed were not relevant to the off-target effects of BRAF and were therefore not retrieved by Causaly. By machine-reading the literature, Causaly separated the signal from the noise, thereby increasing the sensitivity of the search.

Further, 60 additional documents were identified by Causaly, including preprints, patents, and clinical trials. Around 78% (47 of 60) of these articles are relevant and explain the observed off-target effects of BRAF. The remaining 13 papers are useful for identifying underlying mechanisms and pathways linked to the off-target effects of BRAF.

Causaly not only identifies more relevant papers but allows you to delve deeper into the underlying signaling cascades to uncover potential safety signals. By machine-reading the biomedical literature, Causaly extracted 106 off-target effects of BRAF. For example, a pro-apoptotic BCL-2 family members was shown to be upregulated by the inhibition of ERK signaling and BRAF melanoma cells treated with vemurafenib (a BRAF inhibitor).³ The relationship between BRAF and BCL2 proteins can be further explored using Causaly’s hypothesis generation tool, revealing 90 pathways which may mediate the effect of BRAF on BCL2.

This comparison highlights the limitations of traditional keyword searching, which is time-consuming and is not always comprehensive, as demonstrated here with PubMed. In contrast, Causaly’s AI can sift through and remove irrelevant data from biomedical searches, offering a more thorough and accurate understanding of entire research landscapes. This not only streamlines knowledge acquisition but ensures accuracy and precision in navigating the biomedical literature.

Interested in finding out more about how Causaly’s AI can help identify safety red flags early on in drug development?